WO2012047047A2 - Method and apparatus for encoding/decoding video using high-precision filter - Google Patents

Method and apparatus for encoding/decoding video using high-precision filter Download PDF

Info

Publication number
WO2012047047A2
WO2012047047A2 PCT/KR2011/007418 KR2011007418W WO2012047047A2 WO 2012047047 A2 WO2012047047 A2 WO 2012047047A2 KR 2011007418 W KR2011007418 W KR 2011007418W WO 2012047047 A2 WO2012047047 A2 WO 2012047047A2
Authority
WO
WIPO (PCT)
Prior art keywords
block
subsample
image
prediction
value
Prior art date
Application number
PCT/KR2011/007418
Other languages
French (fr)
Korean (ko)
Other versions
WO2012047047A3 (en
Inventor
송진한
임정연
이영렬
문주희
김해광
전병우
한종기
김정필
김대연
Original Assignee
에스케이텔레콤 주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from KR1020110072196A external-priority patent/KR101341993B1/en
Application filed by 에스케이텔레콤 주식회사 filed Critical 에스케이텔레콤 주식회사
Priority to CN201180054874.XA priority Critical patent/CN103210649B/en
Publication of WO2012047047A2 publication Critical patent/WO2012047047A2/en
Publication of WO2012047047A3 publication Critical patent/WO2012047047A3/en
Priority to US13/857,708 priority patent/US9420281B2/en
Priority to US15/031,366 priority patent/US9706222B2/en
Priority to US15/031,378 priority patent/US9602834B2/en
Priority to US15/371,215 priority patent/US10158880B2/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/59—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial sub-sampling or interpolation, e.g. alteration of picture size or resolution
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146—Data rate or code amount at the encoder output
    • H04N19/152—Data rate or code amount at the encoder output by measuring the fullness of the transmission buffer
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding

Definitions

  • An embodiment of the present invention relates to a method and apparatus for image encoding / decoding using a high precision filter. More particularly, the present invention relates to a method and apparatus for generating a prediction signal using a high precision filter to generate a high precision image signal and encoding / decoding an image using the same.
  • Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) have developed video compression techniques that are superior to the existing MPEG-4 Part 2 and H.263 standards.
  • the new standard is called H.264 / AVC (Advanced video Coding), which was jointly released as MPEG-4 Part 10 AVC and ITU-T Recommendation H.264.
  • H.264 / AVC (hereinafter, referred to as H.264)
  • JCT-VC Joint Collaborative Team on Video Coding
  • JCT-VC Joint Collaborative Team on Video Coding
  • the intra prediction encoding method predicts using a prediction value from a block encoded in a frame currently being encoded, and the inter prediction prediction encoding a block of the current frame by estimating a motion from a previously reconstructed frame. The method is used.
  • Intra picture prediction methods using intra luma signals include intra 4 ⁇ 4 prediction, intra 16 ⁇ 16 prediction, intra 8 ⁇ 8 prediction, and the like depending on the prediction direction and the block size to be encoded.
  • 1 is a diagram illustrating nine conventional 4x4 intra prediction modes.
  • intra 4 ⁇ 4 prediction includes a vertical mode, a horizontal mode, a direct current mode, a diagonal down-left mode, and a diagonal down-right.
  • Intra 16 ⁇ 16 prediction there are four prediction modes in the intra 16 ⁇ 16 prediction including a vertical mode, a horizontal mode, a DC mode, and a plane mode.
  • Intra 8 ⁇ 8 prediction also has four prediction modes similar to intra 16 ⁇ 16 prediction.
  • Inter predictive coding of video having a 4: 2: 0 image type uses motion compensation that predicts a current block by dividing an image frame and estimating a motion from a previously encoded frame. If the block size of motion compensation is small, more accurate prediction can be made, but the amount of code increases because the motion vector information must be encoded for each block. Also, when performing motion compensation, not only the motion vector is searched for integer samples with integer pixels but also for subsample positions with resolutions of 1/4 samples for luma components and 1/8 samples for chrominance components. As a result, a more accurate motion vector search method has been used. However, since the luminance and chrominance samples of the subsample position do not exist in the reference picture, these values should be generated by interpolating adjacent integer samples of the reference picture.
  • an embodiment of the present invention uses a high-precision filter that is more precise than linear interpolation in interpolating an image to improve the compression efficiency of the image and to effectively reconstruct the image to improve the subjective image quality.
  • the value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block is FIR.
  • a prediction block for a chrominance component is generated from a value interpolated using a filter and linear interpolation, and the residual block is generated by subtracting the prediction block from the chrominance component of the current block, and the residual block is transformed and quantized to quantize frequency transform.
  • An image encoder for generating a block and encoding the bitstream; And generating a quantized frequency transform block from the bitstream, and inversely quantizing and inverse transforming the quantized frequency transform block to subtract the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block.
  • a prediction block for a color difference component is generated from an interpolated value of a sample component using an FIR filter and linear interpolation, and the color block component of the current block to be restored is restored by adding the reconstructed residual block and the generated prediction block.
  • an image encoding / decoding apparatus comprising an image decoder.
  • an embodiment of the present invention in the apparatus for encoding an image, the value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block referred to by the motion vector of the color difference component of the current block A prediction unit for generating a prediction block for the color difference component from the interpolated values using the FIR filter and linear interpolation; A subtraction unit which subtracts the prediction block from the color difference component of the current block to generate a residual block; A transformer for converting the residual block to generate a frequency transform block; A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And an encoder configured to encode the quantized frequency transform block into a bitstream.
  • an embodiment of the present invention in the apparatus for encoding an image, the integer pixel of the periphery for 1/2 subsample of the reference block referred to by the motion vector of the current block
  • a prediction unit generating a prediction block by obtaining a half-sample high precision applying a filtering coefficient to a value and obtaining values of all subsamples of the reference block by using the half-sample high precision
  • a subtraction unit for generating a residual block by subtracting the prediction block from the current block
  • a transformer for converting the residual block to generate a frequency transform block
  • a quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block
  • an encoder configured to encode the quantized frequency transform block into a bitstream.
  • an embodiment of the present invention to achieve another object of the present invention, an apparatus for decoding an image, the decoding unit for extracting a quantized frequency transform block from the bitstream; An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block; An inverse transform unit which inversely transforms the frequency transform block to restore a residual block; A prediction unit for generating a prediction block for the chrominance component from a value obtained by interpolating a value of the subsample component of the chrominance component of the reference block referred to by the motion vector of the chrominance component of the current block using an FIR filter and linear interpolation; And an adder configured to add the reconstructed residual block and the prediction block to reconstruct the color difference component of the current block.
  • an embodiment of the present invention to achieve another object of the present invention, an apparatus for decoding an image, the decoding unit for extracting a quantized frequency transform block from the bitstream; An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block; An inverse transform unit which inversely transforms the frequency transform block to restore a residual block; For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction unit generating a prediction block by obtaining values of all subsamples of the subsample; And an adder configured to reconstruct the current block by adding the reconstructed residual block and the prediction block.
  • an embodiment of the present invention in the method for encoding / decoding an image, the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block Generate a prediction block for a chrominance component from an interpolated value using a FIR filter and linear interpolation, subtract the prediction block from the chrominance component of the current block to generate a residual block, transform and quantize the residual block Generating an quantized frequency transform block and encoding the bitstream into a bitstream; And generating a quantized frequency transform block from the bitstream, and inversely quantizing and inverse transforming the quantized frequency transform block to subtract the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block.
  • a prediction block for a color difference component is generated from an interpolated value of a sample component using an FIR filter and linear interpolation, and the color block component of the current block to be restored is restored by adding the reconstructed residual block and the generated prediction block. It provides a video encoding / decoding method comprising a video decoding step.
  • the motion vector of the color difference component is obtained by motion compensation of the color difference component of the current block and the subsample component of the color difference component
  • a prediction step of generating a prediction block for a chrominance component from a value obtained by interpolating the value of F using a FIR filter and linear interpolation A subtraction step of generating a residual block by subtracting the prediction block from the color difference component of the current block; Transforming the residual block to generate a frequency transform block;
  • an encoding step of encoding the quantized frequency transform block into a bitstream
  • an embodiment of the present invention provides a method for encoding an image, wherein integer integers of neighboring one-half subsamples of a reference block referred to by a motion vector of the current block
  • a prediction step of generating a prediction block by obtaining a half-sample high precision applying a filtering coefficient to a value and calculating values of all subsamples of the reference block by using the half-sample high precision Subtracting the prediction block from the current block to generate a residual block; Transforming the residual block to generate a frequency transform block; A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And an encoding step of encoding the quantized frequency transform block into a bitstream.
  • an embodiment of the present invention to achieve another object of the present invention, a method for decoding an image, the decoding step of generating a quantized frequency transform block from a bitstream; An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block; An inverse transform step of restoring a residual block by inversely transforming the frequency transform block; Predicting a motion vector of the color difference component by motion compensating the color difference component of the current block and generating a prediction block for the color difference component from values obtained by interpolating values of the subsample components of the color difference component using an FIR filter and linear interpolation; And adding the reconstructed residual block and the prediction block to reconstruct the color difference component of the current block.
  • an embodiment of the present invention to achieve another object of the present invention, a method for decoding an image, the decoding step of extracting a quantized frequency transform block from the bitstream; An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block; An inverse transform step of restoring a residual block by inversely transforming the frequency transform block; For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction step of generating a prediction block by obtaining values of all subsamples of the subsample; And an addition step of reconstructing the current block by adding the reconstructed residual block and the prediction block.
  • the encoding efficiency can be increased by reducing the difference between the actual block and the predicted block, thereby improving the compression efficiency of the current block. It is possible to effectively reconstruct an image by decoding a block converted into a bitstream by improving and considering an encoding method.
  • 1 is a diagram illustrating nine conventional 4x4 intra prediction modes.
  • FIG. 2 is a diagram illustrating four conventional 16x16 intra prediction modes.
  • FIG. 3 is a diagram illustrating motion prediction using a conventional subsample in a luminance (Luma) component.
  • FIR Finite Impulse Response
  • FIG. 6 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.
  • FIG. 7 is an exemplary diagram for describing a process of the prediction unit 610 interpolating a subsample value in a block according to an embodiment of the present invention.
  • FIG. 8 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.
  • FIG. 9 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.
  • FIG. 10 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.
  • a video encoding apparatus (Video Encoding Apparatus), a video decoding apparatus (Video Decoding Apparatus) to be described below is a personal computer (PC), notebook computer, personal digital assistant (PDA), portable multimedia player (PMP) It may be a user terminal such as a portable multimedia player (PSP), a PlayStation Portable (PSP), a wireless communication terminal, a smart phone, a TV, or a server terminal such as an application server or a service server.
  • a communication device such as a communication modem for communicating with various devices or a wired / wireless communication network, a memory for storing various programs and data for encoding or decoding an image or inter or intra prediction for encoding or decoding, and executing a program And a microprocessor for controlling and the like. It can mean a variety of devices.
  • the image encoded in the bitstream by the video encoding apparatus is real-time or non-real-time through the wired or wireless communication network, such as the Internet, local area wireless communication network, wireless LAN network, WiBro network, mobile communication network, or the like, or a cable, universal serial bus (USB: Universal) It may be transmitted to an image decoding apparatus through various communication interfaces such as a serial bus, and may be decoded by the image decoding apparatus to restore and reproduce the image.
  • wired or wireless communication network such as the Internet, local area wireless communication network, wireless LAN network, WiBro network, mobile communication network, or the like, or a cable, universal serial bus (USB: Universal) It may be transmitted to an image decoding apparatus through various communication interfaces such as a serial bus, and may be decoded by the image decoding apparatus to restore and reproduce the image.
  • USB universal serial bus
  • a video may be composed of a series of pictures, and each picture may be divided into a predetermined area such as a frame or a block.
  • the divided blocks may be classified into intra blocks and inter blocks according to an encoding method.
  • An intra block refers to a block that is encoded by using an intra prediction coding scheme. Intra prediction coding is performed by using pixels of blocks that have been previously encoded, decoded, and reconstructed in a current picture that performs current encoding. A prediction block is generated by predicting pixels of a block, and a difference value with pixels of the current block is encoded.
  • An inter block refers to a block that is encoded using inter prediction coding.
  • Inter prediction coding generates a prediction block by predicting a current block in a current picture by referring to one or more past pictures or future pictures, and then generates a current block. This is a method of encoding the difference value with.
  • a frame referred to for encoding or decoding the current picture is referred to as a reference frame.
  • FIG. 3 is a diagram illustrating motion prediction using a conventional subsample in a luminance (Luma) component.
  • a motion vector may be obtained from an integer sample unit up to a subsample position of 1/4 samples.
  • FIR Finite Impulse Response
  • pixel values at 1/2 sample positions are interpolated using 6 values having integer pixel values and 6 filter coefficients ⁇ 1, -5, 20, 20, -5, 1 ⁇ and 1
  • the components of the / 4 sample are interpolated with interpolated 1/2 sample pixel values and integer pixel values or linear interpolation using two 1/2 sample pixel values to interpolate pixel values at positions of 1/4 samples as shown in the following equation. do.
  • FIG. 5 is a diagram illustrating linear interpolation for a chroma sample.
  • multiplying weight values by considering bi-linear interpolation using bi-linear interpolation using the values of four integer pixels as in the following example shows that 1 / The pixel values of the 8 subsamples are interpolated.
  • ⁇ a [(6 ⁇ 5 ⁇ A) + (2 ⁇ 5 ⁇ B) + (6 ⁇ 3 ⁇ C) + (2 ⁇ 3 ⁇ D)] / 64 ⁇ .
  • the color difference signal is 1/4 (horizontal 1/2, vertical 1/2) of the luminance signal in the 4: 2: 0 image form.
  • FIG. 6 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.
  • the image encoding apparatus 600 encodes an image by generating a subsample prediction value of a chrominance component using a motion vector value of a luminance component of a current block of the image.
  • the prediction unit 610 may include a predictor 610, a subtractor 620, a transformer 630, a quantizer 640, and an encoder 650.
  • the input image to be encoded is input in units of macro blocks.
  • the macro blocks are in the form of M ⁇ N, M and N each have a size of 2 n , and M and N may be the same or different. have.
  • the prediction unit 610 generates a prediction block for the color difference component from the value obtained by interpolating a value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block using an FIR filter and linear interpolation. do.
  • the prediction unit 610 may generate a prediction block using another frame to predict the current macro block. That is, the prediction unit 610 may generate a motion vector through motion estimation in a previous frame that has already been encoded and reconstructed, and generate a prediction block in a motion compensation process using the motion vector. In this case, the prediction unit 610 may use the same motion vector value even in the color difference component by using the motion vector value of the luminance component, and use the high-precision FIR to determine the value of the subsample component of the block in the reference frame indicated by the motion vector. A prediction block having prediction values using a filter and linear interpolation may be generated.
  • the subtractor 620 generates a residual signal by calculating a difference value between the original pixel value of each pixel of the current block and the predicted value generated by the predictor 610.
  • the converter 630 converts the residual signal generated by the subtractor 620 into the frequency domain.
  • the transform unit 630 uses various transformation techniques for transforming an image signal of a time axis into a frequency axis, such as Discrete Cosine Transform (DCT) or Wavelet Transform.
  • DCT Discrete Cosine Transform
  • Wavelet Transform Discrete Cosine Transform
  • the quantization unit 640 quantizes the frequency conversion block formed of the residual signal converted into the frequency domain by the conversion unit 630.
  • various quantization techniques such as dead zone uniform threshold quantization (DZUTQ) or quantization weighted matrix (Quantization Weighted Matrix) may be used.
  • the encoder 650 encodes the quantized frequency transform block including the frequency coefficient quantized by the quantizer 640 into a bitstream.
  • an entropy encoding technique may be used, but various encoding techniques may be used without being limited thereto.
  • the encoder 650 may include not only a bit string encoding the quantized frequency coefficients but also various pieces of information necessary to decode the encoded bit string in the encoded data. That is, the coded data includes a field including a coded block pattern (CBP), a delta quantization parameter, and a bit string in which the quantization frequency coefficients are encoded, and information necessary for prediction (eg, intra prediction). In the case of Intra prediction mode or a motion vector in the case of inter prediction, etc.) may include a field that contains a bit.
  • CBP coded block pattern
  • a delta quantization parameter e.g., a bit string in which the quantization frequency coefficients are encoded
  • information necessary for prediction eg, intra prediction
  • Intra prediction mode or a motion vector in the case of inter prediction, etc. may include a field that contains a bit.
  • the inverse quantization unit 660 inverse quantizes the transformed and quantized residual block (ie, the quantized frequency transform block), and the inverse transform unit 670 inverses the inverse quantized transform residual block. To reconstruct the residual block.
  • inverse quantization and inverse transformation may be performed by inversely performing a transformation process performed by the transform unit 630 and a quantization process performed by the quantization unit 640. That is, the inverse quantization unit 660 and the inverse transform unit 670 are information on transform and quantization generated and transmitted from the transform unit 630 and the quantization unit 640 (for example, information on a transform and quantization type). Inverse quantization and inverse transformation may be performed using.
  • the adder 680 adds the prediction block generated by the predictor 610 and the residual block generated by the inverse transform unit 670 to generate a reconstructed block.
  • the frame memory 690 stores a block reconstructed by the adder 680 and is used as a reference block to generate a prediction block when performing intra or inter prediction.
  • FIG. 7 is an exemplary diagram for describing a process of the prediction unit 610 interpolating a subsample value in a block according to an embodiment of the present invention.
  • Interpolation of the subsample of FIG. 7 may be performed using a method such as Equation 1 to Equation 4.
  • the 1/2 subsample value can be obtained by multiplying the integer pixel values by a predetermined value and then using the 1/2 sample amplification value which is the sum of these multiplied values.
  • the 1/2 subsample value can be obtained by using three left integer pixel values A, B, and C and three right integer pixel values D, E, and F.
  • the 1/4 subsample value can be obtained using the nearest integer pixel value C and the half sample amplification value d '.
  • the formula for obtaining b is obtained by dividing the form of linear interpolation between the nearest integer pixel value (C) and the half-sample amplification value (d ') instead of directly linearly interpolating the 1/2 subsample. The error due to the integerization can be eliminated.
  • b ' can also be obtained using the nearest integer pixel value (C) and the half-sample amplification value (d'), so that the 1/8 subsample value is also the nearest integer pixel value (C) and half-sample amplification. Can be obtained using the value d '.
  • Equations 1, 2, 3, 4, and 7 A, B, C, D, E, and F represent integer pixel values of the color difference component, d represents 1/2 subsample of the color difference component, and b represents the color difference component. 1/4 subsample, a, represents 1/8 subsample of the color difference component.
  • the value of the 1/2 subsample of the chrominance component is generated by using a high precision filter (where the high precision filter can use various filters such as FIR filter), and as shown in Equation 2 Similarly, the values of quarter subsamples use high precision filters and linear interpolation.
  • high-precision linear interpolation can be performed using integer pixels and 1/4 subsamples as shown in Equation 3, and also integer pixels and 1/2 subsamples are used as shown in Equation 4. High accuracy linear interpolation can be achieved.
  • w1 and w2 represent weight values multiplied by integer pixels and 1/2 subsample.
  • the 1/2 subsample value (e.g., d) using Equation 1 and the 1/4 subsample value (e.g., b) using Equation 2 are generated using a high-precision FIR filter, and Equation 3 is used. Since a 1/8 subsample value (eg, a) is generated through linear interpolation between two pixel values, all subsample values can be generated more accurately than when interpolation is performed using only linear interpolation.
  • Equation 5 is an equation in which the above rounding operation is added to Equation 3 below.
  • the FIR filter is a type of digital filter and performs filtering with only certain values of the input signal. Therefore, for the FIR filter, the impulse response, which is the characteristic function of the filter, has a finite length. In addition, in the form of the FIR filter, it does not have a feedback component, so the execution time increases due to the higher order when implementing the same characteristic, but the FIR filter is important when phase shift (ie, the shape maintenance of the waveform between the input and the output) is important. Can be used.
  • the high precision FIR filter used here is a 1/2 subsample using an FIR filter, and a 1/4 subsample is 1/2 divided by 32 to the value (d ') using the FIR filter as shown in Equation 2.
  • the subsample value (d in Equation 1) multiply the value before dividing by 32 (that is, the FIR filtering result d 'obtained by obtaining the value of 1/2 subsample) and the integer pixel value by 32.
  • Linear interpolation of the value 32XC prevents the loss of information due to the division operation in the middle and thus interpolates more precise values.
  • the prediction block for the current block may be obtained as sample values.
  • the number of bits per sample of the prediction block may be better than the number of bits per sample of the current block.
  • Equation 4 only the value obtained by interpolating a (1/8 subsample) is described. However, the weight (w1, w2) is set differently from the case of obtaining a even when obtaining b (1/4 subsample). B can be obtained.
  • f can be obtained by using f 'instead of b', f instead of b, and D instead of C in Equation 2, and g 'instead of a', g 'instead of a, and f' instead of b 'in Equation 3
  • 1/8 subsample c can be obtained by interpolating b and d
  • e can be obtained by interpolating d and f, so that the nearest integer pixel value (C) and 1/2 sample are similar to a or g. It can be obtained using the amplification value d '.
  • the surrounding integer pixel value is obtained to obtain 1/2 subsample, and the value of the finer subsample (1/4 subsample, 1/8 subsample, etc.) is smaller than the left and right (or the top and bottom) subtle subsamples.
  • Interpolation i.e., using 1/2 subsample and integer pixel to obtain a 1/4 sample value
  • the case of using the value of the subsample or the integer pixel for interpolation of the subsamples of various positions may be inferred by those skilled in the art even if the case of all the subsamples is not described.
  • FIR filters having 6-tab or more for 1/2 samples, and high-precision functions such as Equation 2 above for 1/4 samples It is important to interpolate to have.
  • the reference blocks in the reference frame may be interpolated using Equations 1 to 3, and in the case of the 4: 4: 4 picture form, the luminance component in the reference block Since the magnitudes of the chrominance components are the same, the chrominance components, like the luminance component, may be interpolated to a quarter sample position. Therefore, if the 1/8 sample position interpolation of Equation 3 is omitted, it can be used in 4: 4: 4 image form.
  • the horizontal color difference component is interpolated to the 1/8 subsample position as shown in Equations 1 to 3 below.
  • the subsample may be generated by interpolating up to the position of the 1/4 subsample using Equations 1 and 2 similarly to the 4: 4: 4 image form.
  • TMuC Test Model under Consideration
  • TMuC Test Model under Consideration
  • the chrominance components can be up to 1/16 subsamples in 4: 2: 0 video format. You can interpolate.
  • up to 1/8 sub-positions can be made as stated in one embodiment of the present invention and the positions of 1/16 subsamples can be interpolated again using linear interpolation.
  • the filtering and interpolation methods have been described with reference to color difference components as an example.
  • the method may be applied to various blocks such as luminance components and blocks of R, G, and B colors as well as color difference components.
  • the same can be applied to various blocks such as luminance components and blocks of R, G, and B colors as well as color difference components in the decoding method described later.
  • FIG. 8 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.
  • the image decoding apparatus 800 is an apparatus for generating and decoding a subsample prediction value of a chrominance component using a motion vector value of a luminance component of a current block of an image. And a decoder 810, an inverse quantizer 820, an inverse transformer 830, an adder 840, and a predictor 850.
  • the decoder 810 extracts a quantized frequency transform block by decoding the bitstream.
  • the decoder 810 may decode and extract not only the quantized frequency transform block but also information necessary for decoding by decoding the encoded data.
  • Information necessary for decoding refers to information necessary for decoding the coded bit string in the encoded data (ie, the bitstream). For example, information about a block type, information about a motion vector, information about a transform and quantization type, and the like. It can be, and a variety of other information can be.
  • the decoder 810 decodes a bitstream, which is data encoded by the image encoding apparatus 600, extracts a quantized frequency transform block including pixel information of the current block of the image, and extracts the extracted prediction signal.
  • the necessary information is transmitted to the prediction unit 850.
  • the prediction unit 850 may predict the current block in the same manner as the prediction unit 610 of the image encoding apparatus 600 by using the information necessary for the prediction transmitted from the decoder 810.
  • the prediction unit 850 generates a prediction block for the color difference component from the value obtained by interpolating the values of the subsample components of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block using an FIR filter and linear interpolation. do.
  • the motion vector value of the luminance component reconstructed in the bitstream is used, and the integer pixel value of the chrominance component of the reference block referred to by the motion vector is converted into a high precision FIR filter.
  • the prediction unit 850 of the image decoding apparatus 800 generates a subsample in the same manner as the prediction unit 610 of the image encoding apparatus 600 described above with reference to FIG. 6. Therefore, detailed description thereof will be omitted to avoid redundant description.
  • the inverse quantizer 820 inverse quantizes the quantized frequency transform block extracted from the bitstream by the decoder 810.
  • the inverse transformer 830 inversely transforms the frequency transform block inversely quantized by the inverse quantizer 820 into the time domain.
  • the adder 840 reconstructs the original pixel value of the current block by adding the residual pixel restored by the inverse transform by the inverse transformer 830 and the predicted pixel value by the predictor 850.
  • the current block reconstructed by the adder 840 may be transferred to the frame memory 860 and used by the predictor 850 to predict another block.
  • the frame memory 860 stores the reconstructed image to enable generation of the intra prediction block and the inter prediction block.
  • An image encoding / decoding apparatus may be configured by connecting a bitstream output terminal of the image encoding apparatus 600 of FIG. 6 to a bitstream input terminal of the image decoding apparatus 800 of FIG. 8.
  • a value obtained by interpolating a value of a subsample component of a chrominance component of a reference block referred to by a motion vector of a chrominance component of a current block using an FIR filter and linear interpolation An image of generating a prediction block for a chrominance component from the chrominance component, subtracting the prediction block from the chrominance component of the current block, generating a residual block, transforming and quantizing the residual block, and generating a quantized frequency transform block and encoding the bitstream
  • a quantized frequency transform block is generated from an encoder and a bitstream, and the quantized frequency transform block is inversely quantized and inverse transformed so that the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block.
  • a video decoder for generating a prediction block for the chrominance component from and reconstructing the chrominance component of the current block to be reconstructed by adding the reconstructed residual block and the generated prediction block.
  • the image encoder may be implemented by the image encoding apparatus 600 according to an embodiment of the present invention, and the image decoder may be implemented by the image decoding apparatus 800 according to the embodiment of the present invention.
  • FIG. 9 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.
  • the image encoding apparatus 600 generates a subsample prediction value of the color difference component using the motion vector value of the luminance component of the current block of the image (S910), and the difference between the original pixel value and the prediction pixel value of the current block.
  • a subtraction step (S920) of generating a residual signal by calculating a value a conversion step (S930) of converting the generated residual signal into a frequency domain using a DCT transform or a wavelet transform, and quantizing the residual signal transformed into a frequency domain
  • the image is encoded through a quantization step S940 and an encoding step S950 of encoding the quantized frequency transform residual signal into a bitstream.
  • the prediction step (S910) is a function of the prediction unit 610
  • the subtraction step (S920) is a function of the subtraction unit 620
  • the transform step (S930) is a function of the transformer 630
  • FIG. 10 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.
  • the video decoding apparatus 800 which receives and stores a bitstream of an image through a wired or wireless communication network or a cable, stores the current block of the image in order to reproduce the image according to a user's selection or an algorithm of another program being executed.
  • a subsample prediction value of the chrominance component is generated using the motion vector value, and the image is decoded and reconstructed.
  • the image decoding apparatus 800 decodes the bitstream and extracts a quantized frequency transform residual signal representing information about the pixel value of the current block of the image (S1010).
  • An inverse quantization step of inversely quantizing a quantized frequency transform residual signal, an inverse transform step of inversely transforming an inverse quantized frequency transform residual signal into a time domain, and a current block represented by a residual signal inversely transformed into a time domain and restored A prediction step (S1040) of generating a subsample prediction value of a chrominance component using the motion vector value of the luminance component of the predicted value of, and prediction of each pixel of the current block predicted in step S1040 and the residual signal of the current block restored in step S1040.
  • the bitstream transmitted through the addition step S1050 of restoring the original pixel value of the current block by adding the pixel value is decoded.
  • the decoding step (S1010) corresponds to the operation of the decoding unit 810
  • the inverse quantization step (S1020) corresponds to the operation of the inverse quantization unit 820
  • the inverse transform step (S1030) of the inverse transformer (830) corresponds to the prediction step (S1040)
  • the addition step (S1050) corresponds to the operation of the adding unit 840, and thus, detailed description thereof will be omitted.
  • An image encoding / decoding method may be realized by combining the image encoding method according to an embodiment of the present invention and the image decoding method according to an embodiment of the present invention.
  • a value obtained by interpolating a value of a subsample component of a chrominance component of a reference block referred to by a motion vector of a chrominance component of a current block using an FIR filter and linear interpolation An image of generating a prediction block for a chrominance component from the chrominance component, subtracting the prediction block from the chrominance component of the current block, generating a residual block, transforming and quantizing the residual block, and generating a quantized frequency transform block and encoding the bitstream
  • a color difference component of a reference block referred to by a motion vector of a color difference component of a current block to generate and quantize the quantized frequency transform block from the encoding step and the bitstream, and to inversely quantize and inverse transform the quantized frequency transform block.
  • the subsample components of are interpolated using FIR filter and linear interpolation. And generating a prediction block for the color difference component from the value, and adding the reconstructed residual block and the generated prediction block to reconstruct the color difference component of the current block to be reconstructed.
  • the image encoding step may be implemented as an image encoding step according to an embodiment of the present invention
  • the image decoding step may be implemented as an image decoding step according to an embodiment of the present invention.
  • all the components constituting the embodiment of the present invention may be implemented in one independent hardware, each of some or all of the components are selectively combined to some or all of the functions combined in one or a plurality of hardware. It may be implemented as a computer program having a program module for performing the operation. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention.
  • the storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.
  • encoding is applied to an interpolation that minimizes a difference between an actual component and a predicted component and encoding / decoding of the image using the encoding.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

One embodiment of the present invention relates to a method and an apparatus for encoding/decoding video using a high-precision filter. One embodiment of the present invention provides the method and the apparatus for encoding/decoding the video, and in the method for encoding/decoding the video, the value of a subsample element of a reference block, which is referenced by the motion vector of a current block, is encoded and decoded by generating a predictive block from a value that is interpolated using an FIR filter and linear interpolation, so as to more accurately interpolate the current block, thereby increasing encoding efficiency by reducing the difference between the actual block and the predictive block, and thus improving compression efficiency.

Description

고정밀 필터를 이용한 영상 부호화/복호화 방법 및 장치Method and apparatus for image coding / decoding using high precision filter
본 발명의 실시예는 고정밀 필터를 이용한 영상 부호화/복호화 방법 및 장치에 관한 것이다. 더욱 상세하게는, 고정밀도의 영상 신호를 생성하기 위하여 고정밀 필터를 이용하여 예측신호를 생성하고 이를 이용한 영상의 부호화/복호화 방법 및 장치에 관한 것이다.An embodiment of the present invention relates to a method and apparatus for image encoding / decoding using a high precision filter. More particularly, the present invention relates to a method and apparatus for generating a prediction signal using a high precision filter to generate a high precision image signal and encoding / decoding an image using the same.
이 부분에 기술된 내용은 단순히 본 발명의 실시예에 대한 배경 정보를 제공할 뿐 반드시 종래기술을 구성하는 것은 아니다.The content described in this section merely provides background information on the embodiments of the present invention and does not necessarily constitute prior art.
MPEG(Moving Picture Experts Group)과 VCEG(Video Coding Experts Group)은 기존의 MPEG-4 Part 2와 H.263 표준안보다 우수하고 뛰어난 비디오 압축 기술을 개발하였다. 이 새로운 표준안은 H.264/AVC(Advanced video Coding)이라 하며, MPEG-4 Part 10 AVC와 ITU-T Recommendation H.264로 공동으로 발표되었다. 이러한 H.264/AVC(이하 'H.264'라 약칭함)에서는 다양한 인코딩 방법을 사용하여 화질과 성능 향상에서 많은 발전을 가져왔다. 또한 현재 JCT-VC(Joint Collaborative Team on Video Coding)라는 MPEG과 VCEG의 공동 팀에 의하여 HD(High-Definition)급 이상의 화질에 대한 새로운 표준안에 대하여 표준화 회의가 진행 중이다Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) have developed video compression techniques that are superior to the existing MPEG-4 Part 2 and H.263 standards. The new standard is called H.264 / AVC (Advanced video Coding), which was jointly released as MPEG-4 Part 10 AVC and ITU-T Recommendation H.264. In H.264 / AVC (hereinafter, referred to as H.264), various encoding methods have been used to improve the image quality and performance. In addition, a joint meeting between MPEG and VCEG, called Joint Collaborative Team on Video Coding (JCT-VC), is currently conducting a standardization conference on a new standard for high-definition (HD) or higher quality.
기존의 동영상 부호화 방법에서는 현재 부호화중인 프레임 내에서 부호화된 블록으로부터 예측 값을 사용하여 예측하는 화면 내 예측 부호화 방법과 이전에 복원된 프레임으로부터 움직임을 추정하여 현재 프레임의 블록을 예측하는 화면 간 예측 부호화 방법이 사용되고 있다.In the existing video encoding method, the intra prediction encoding method predicts using a prediction value from a block encoded in a frame currently being encoded, and the inter prediction prediction encoding a block of the current frame by estimating a motion from a previously reconstructed frame. The method is used.
휘도(Luma) 신호에서 화면 내 예측 방법에는 예측 방향과 부호화되는 블록 사이즈에 따라 인트라 4×4 예측, 인트라 16×16 예측, 인트라 8×8 예측 등이 사용되고 있다.Intra picture prediction methods using intra luma signals include intra 4 × 4 prediction, intra 16 × 16 prediction, intra 8 × 8 prediction, and the like depending on the prediction direction and the block size to be encoded.
도 1은 통상적인 9 가지의 4×4 인트라 예측 모드를 나타낸 도면이다.1 is a diagram illustrating nine conventional 4x4 intra prediction modes.
도 1을 참조하면, 인트라 4×4 예측에는, 수직 (Vertical) 모드, 수평 (Horizontal) 모드, DC (Direct Current) 모드, 대각선 왼쪽 (Diagonal down-left) 모드, 대각선 오른쪽 (Diagonal down-right) 모드, 수직 오른쪽 (Vertical-right) 모드, 수평 아래쪽 (Horizontal-down)모드, 수직 왼쪽 (Vertical-left) 및 수평 위쪽 (Horizontal-up)모드를 포함하는 9가지의 예측 모드가 있다.Referring to FIG. 1, intra 4 × 4 prediction includes a vertical mode, a horizontal mode, a direct current mode, a diagonal down-left mode, and a diagonal down-right. There are nine prediction modes, including mode, vertical-right mode, horizontal-down mode, vertical-left, and horizontal-up mode.
도 2는 통상적인 4 가지의 16×16 인트라 예측 모드를 나타낸 도면이다.2 shows four conventional 16x16 intra prediction modes.
도 2를 참조하면, 인트라 16×16 예측에는, 수직 (Vertical) 모드, 수평 (Horizontal) 모드, DC 모드 및 플레인 (Plane)모드를 포함하는 4 가지의 예측 모드가 있다. 인트라 8×8 예측도 인트라 16×16 예측과 유사하게 4 가지의 예측 모드가 있다.Referring to FIG. 2, there are four prediction modes in the intra 16 × 16 prediction including a vertical mode, a horizontal mode, a DC mode, and a plane mode. Intra 8 × 8 prediction also has four prediction modes similar to intra 16 × 16 prediction.
4:2:0 영상 형태를 갖는 비디오의 화면 간 예측 방법(Inter predictive coding)에는 영상 프레임을 분할해서 이전에 부호화된 프레임으로부터 움직임을 추정하여 현재의 블록을 예측하는 움직임보상을 사용하고 있다. 움직임보상의 블록크기를 작게 한다면 보다 정확한 예측을 할 수 있지만 블록마다 움직임벡터 정보를 부호화해야 하기 때문에 부호량이 증가하게 된다. 또한 움직임보상을 할 때 정수 화소를 가지는 정수 샘플에서만 움직임 벡터를 찾는 것이 아니라 휘도(Luma) 성분에 대해 1/4 샘플의 해상도, 색차 성분에 대해 1/8 샘플의 해상도를 가지는 서브 샘플 위치까지 찾음으로써 더욱 정확한 움직임 벡터를 찾는 방법이 사용되고 있다. 하지만 서브 샘플 위치의 휘도와 색차 샘플은 참조 픽쳐 내에 존재하지 않으므로 참조 픽쳐의 인접해 있는 정수 샘플들을 보간하여 이 값들을 생성해 내야 한다.Inter predictive coding of video having a 4: 2: 0 image type uses motion compensation that predicts a current block by dividing an image frame and estimating a motion from a previously encoded frame. If the block size of motion compensation is small, more accurate prediction can be made, but the amount of code increases because the motion vector information must be encoded for each block. Also, when performing motion compensation, not only the motion vector is searched for integer samples with integer pixels but also for subsample positions with resolutions of 1/4 samples for luma components and 1/8 samples for chrominance components. As a result, a more accurate motion vector search method has been used. However, since the luminance and chrominance samples of the subsample position do not exist in the reference picture, these values should be generated by interpolating adjacent integer samples of the reference picture.
이러한 문제점을 해결하기 위해 본 발명의 실시예는, 영상을 보간 하는데 있어서 선형 보간법보다 더욱 정밀한 고 정밀도 필터를 사용함으로 영상의 압축 효율을 향상시키고 영상을 효과적으로 복원하여 주관적 화질을 향상시키는 데 주된 목적이 있다.In order to solve this problem, an embodiment of the present invention uses a high-precision filter that is more precise than linear interpolation in interpolating an image to improve the compression efficiency of the image and to effectively reconstruct the image to improve the subjective image quality. have.
전술한 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화/복호화하는 장치에 있어서, 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화기; 및 비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록의 색차성분을 복원하는 영상 복호화기를 포함하는 것을 특징으로 하는 영상 부호화/복호화 장치를 제공한다.In order to achieve the above object, according to an embodiment of the present invention, in an apparatus for encoding / decoding an image, the value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block is FIR. A prediction block for a chrominance component is generated from a value interpolated using a filter and linear interpolation, and the residual block is generated by subtracting the prediction block from the chrominance component of the current block, and the residual block is transformed and quantized to quantize frequency transform. An image encoder for generating a block and encoding the bitstream; And generating a quantized frequency transform block from the bitstream, and inversely quantizing and inverse transforming the quantized frequency transform block to subtract the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block. A prediction block for a color difference component is generated from an interpolated value of a sample component using an FIR filter and linear interpolation, and the color block component of the current block to be restored is restored by adding the reconstructed residual block and the generated prediction block. Provided is an image encoding / decoding apparatus comprising an image decoder.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화하는 장치에 있어서, 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하는 예측부; 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산부; 상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환부; 상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화부; 및 상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화부를 포함하는 것을 특징으로 하는 영상 부호화 장치를 제공한다.In addition, to achieve another object of the present invention, an embodiment of the present invention, in the apparatus for encoding an image, the value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block A prediction unit for generating a prediction block for the color difference component from the interpolated values using the FIR filter and linear interpolation; A subtraction unit which subtracts the prediction block from the color difference component of the current block to generate a residual block; A transformer for converting the residual block to generate a frequency transform block; A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And an encoder configured to encode the quantized frequency transform block into a bitstream.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화하는 장치에 있어서, 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측부; 상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산부; 상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환부; 상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화부; 및 상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화부를 포함하는 것을 특징으로 하는 영상 부호화 장치를 제공한다.In addition, in order to achieve another object of the present invention, an embodiment of the present invention, in the apparatus for encoding an image, the integer pixel of the periphery for 1/2 subsample of the reference block referred to by the motion vector of the current block A prediction unit generating a prediction block by obtaining a half-sample high precision applying a filtering coefficient to a value and obtaining values of all subsamples of the reference block by using the half-sample high precision; A subtraction unit for generating a residual block by subtracting the prediction block from the current block; A transformer for converting the residual block to generate a frequency transform block; A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And an encoder configured to encode the quantized frequency transform block into a bitstream.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 복호화하는 장치에 있어서, 비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화부; 상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화부; 상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환부; 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하는 예측부; 및 상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록의 색차성분을 복원하는 가산부를 포함하는 것을 특징으로 하는 영상 복호화 장치를 제공한다.In addition, an embodiment of the present invention to achieve another object of the present invention, an apparatus for decoding an image, the decoding unit for extracting a quantized frequency transform block from the bitstream; An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block; An inverse transform unit which inversely transforms the frequency transform block to restore a residual block; A prediction unit for generating a prediction block for the chrominance component from a value obtained by interpolating a value of the subsample component of the chrominance component of the reference block referred to by the motion vector of the chrominance component of the current block using an FIR filter and linear interpolation; And an adder configured to add the reconstructed residual block and the prediction block to reconstruct the color difference component of the current block.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 복호화하는 장치에 있어서, 비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화부; 상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화부; 상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환부; 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측부; 및 상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산부를 포함하는 것을 특징으로 하는 영상 복호화 장치를 제공한다.In addition, an embodiment of the present invention to achieve another object of the present invention, an apparatus for decoding an image, the decoding unit for extracting a quantized frequency transform block from the bitstream; An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block; An inverse transform unit which inversely transforms the frequency transform block to restore a residual block; For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction unit generating a prediction block by obtaining values of all subsamples of the subsample; And an adder configured to reconstruct the current block by adding the reconstructed residual block and the prediction block.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화/복호화하는 방법에 있어서, 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화단계; 및 비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록의 색차성분을 복원하는 영상 복호화단계를 포함하는 것을 특징으로 하는 영상 부호화/복호화 방법을 제공한다.Further, in order to achieve another object of the present invention, an embodiment of the present invention, in the method for encoding / decoding an image, the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block Generate a prediction block for a chrominance component from an interpolated value using a FIR filter and linear interpolation, subtract the prediction block from the chrominance component of the current block to generate a residual block, transform and quantize the residual block Generating an quantized frequency transform block and encoding the bitstream into a bitstream; And generating a quantized frequency transform block from the bitstream, and inversely quantizing and inverse transforming the quantized frequency transform block to subtract the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block. A prediction block for a color difference component is generated from an interpolated value of a sample component using an FIR filter and linear interpolation, and the color block component of the current block to be restored is restored by adding the reconstructed residual block and the generated prediction block. It provides a video encoding / decoding method comprising a video decoding step.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화하는 방법에 있어서, 현재 블록의 색차 성분을 움직임 보상하여 색차 성분의 움직임 벡터를 구하고 상기 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하는 예측단계; 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산단계; 상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환단계; 상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화단계; 및 상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화단계를 포함하는 것을 특징으로 하는 영상 부호화 방법을 제공한다.In addition, to achieve another object of the present invention, an embodiment of the present invention, in a method of encoding an image, the motion vector of the color difference component is obtained by motion compensation of the color difference component of the current block and the subsample component of the color difference component A prediction step of generating a prediction block for a chrominance component from a value obtained by interpolating the value of F using a FIR filter and linear interpolation; A subtraction step of generating a residual block by subtracting the prediction block from the color difference component of the current block; Transforming the residual block to generate a frequency transform block; A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And an encoding step of encoding the quantized frequency transform block into a bitstream.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 부호화하는 방법에 있어서, 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측단계; 상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산단계; 상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환단계; 상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화단계; 및 상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화단계를 포함하는 것을 특징으로 하는 영상 부호화 방법을 제공한다.Further, in order to achieve another object of the present invention, an embodiment of the present invention provides a method for encoding an image, wherein integer integers of neighboring one-half subsamples of a reference block referred to by a motion vector of the current block A prediction step of generating a prediction block by obtaining a half-sample high precision applying a filtering coefficient to a value and calculating values of all subsamples of the reference block by using the half-sample high precision; Subtracting the prediction block from the current block to generate a residual block; Transforming the residual block to generate a frequency transform block; A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And an encoding step of encoding the quantized frequency transform block into a bitstream.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 복호화하는 방법에 있어서, 비트스트림으로부터 양자화된 주파수변환 블록을 생성하는 복호화단계; 상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화단계; 상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환단계; 현재 블록의 색차 성분을 움직임 보상하여 색차 성분의 움직임 벡터를 구하고 상기 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하는 예측단계; 및 상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록의 색차성분을 복원하는 가산단계를 포함하는 것을 특징으로 하는 영상 복호화 방법을 제공한다.In addition, an embodiment of the present invention to achieve another object of the present invention, a method for decoding an image, the decoding step of generating a quantized frequency transform block from a bitstream; An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block; An inverse transform step of restoring a residual block by inversely transforming the frequency transform block; Predicting a motion vector of the color difference component by motion compensating the color difference component of the current block and generating a prediction block for the color difference component from values obtained by interpolating values of the subsample components of the color difference component using an FIR filter and linear interpolation; And adding the reconstructed residual block and the prediction block to reconstruct the color difference component of the current block.
또한, 본 발명의 다른 목적을 달성하기 위해 본 발명의 일 실시예는, 영상을 복호화하는 방법에 있어서, 비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화단계; 상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화단계; 상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환단계; 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측단계; 및 상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산단계를 포함하는 것을 특징으로 하는 영상 복호화 방법을 제공한다.In addition, an embodiment of the present invention to achieve another object of the present invention, a method for decoding an image, the decoding step of extracting a quantized frequency transform block from the bitstream; An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block; An inverse transform step of restoring a residual block by inversely transforming the frequency transform block; For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction step of generating a prediction block by obtaining values of all subsamples of the subsample; And an addition step of reconstructing the current block by adding the reconstructed residual block and the prediction block.
이상에서 설명한 바와 같이 본 발명의 실시예에 의하면, 부호화하고자 하는 현재 블록을 더욱 정확하게 보간함으로써, 실제 블록과 예측된 블록 간의 차이를 줄여 부호화의 효율을 높일 수 있으며, 그를 통해 현재 블록의 압축효율을 향상시키고 부호화방법을 고려하여 비트스트림으로 변환된 블록을 복호화함으로써 효과적으로 영상을 복원하는 효과가 있다.As described above, according to the embodiment of the present invention, by interpolating the current block to be encoded more accurately, the encoding efficiency can be increased by reducing the difference between the actual block and the predicted block, thereby improving the compression efficiency of the current block. It is possible to effectively reconstruct an image by decoding a block converted into a bitstream by improving and considering an encoding method.
도 1은 통상적인 9 가지의 4x4 인트라 예측 모드를 나타낸 도면이다.1 is a diagram illustrating nine conventional 4x4 intra prediction modes.
도 2는 통상적인 4 가지의 16x16 인트라 예측 모드를 나타낸 도면이다.2 is a diagram illustrating four conventional 16x16 intra prediction modes.
도 3은 휘도(Luma) 성분에서 통상적인 서브 샘플을 이용한 움직임 예측을 나타낸 도면이다.3 is a diagram illustrating motion prediction using a conventional subsample in a luminance (Luma) component.
도 4는 통상적인 6-tab FIR(Finite Impulse Response) 필터의 예시이다4 is an illustration of a typical 6-tab Finite Impulse Response (FIR) filter.
도 5는 통상적인 색차(Chroma) 샘플에 대한 선형 보간을 예시한 도면이다.5 illustrates linear interpolation for a typical Chroma sample.
도 6은 본 발명의 일 실시예에 따른 영상 부호화 장치를 간략하게 나타낸 블록 구성도이다.6 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.
도 7은 본 발명의 일 실시예에 따라 예측부(610)가 블록 내의 서브샘플값을 보간하는 과정을 설명하기 위한 예시도이다.7 is an exemplary diagram for describing a process of the prediction unit 610 interpolating a subsample value in a block according to an embodiment of the present invention.
도 8은 본 발명의 일 실시예에 따른 영상 복호화 장치의 구성을 나타낸 블록 구성도이다.8 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.
도 9는 본 발명의 일 실시예에 따른 영상 부호화 방법을 설명하기 위한 순서도이다.9 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.
도 10은 본 발명의 일 실시예에 따른 영상 복호화 방법을 설명하기 위한 순서도이다.10 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.
이하에서 후술할 영상 부호화 장치(Video Encoding Apparatus), 영상 복호화 장치(Video Decoding Apparatus)는 개인용 컴퓨터(PC: Personal Computer), 노트북 컴퓨터, 개인 휴대 단말기(PDA: Personal Digital Assistant), 휴대형 멀티미디어 플레이어(PMP: Portable Multimedia Player), 플레이스테이션 포터블(PSP: PlayStation Portable), 무선 통신 단말기(Wireless Communication Terminal), 스마트폰(Smart Phone), TV 등과 같은 사용자 단말기이거나 응용 서버와 서비스 서버 등 서버 단말기일 수 있으며, 각종 기기 또는 유무선 통신망과 통신을 수행하기 위한 통신 모뎀 등의 통신 장치, 영상을 부호화하거나 복호화하거나 부호화 또는 복호화를 위해 인터 또는 인트라 예측하기 위한 각종 프로그램과 데이터를 저장하기 위한 메모리, 프로그램을 실행하여 연산 및 제어하기 위한 마이크로프로세서 등을 구비하는 다양한 장치를 의미할 수 있다.A video encoding apparatus (Video Encoding Apparatus), a video decoding apparatus (Video Decoding Apparatus) to be described below is a personal computer (PC), notebook computer, personal digital assistant (PDA), portable multimedia player (PMP) It may be a user terminal such as a portable multimedia player (PSP), a PlayStation Portable (PSP), a wireless communication terminal, a smart phone, a TV, or a server terminal such as an application server or a service server. A communication device such as a communication modem for communicating with various devices or a wired / wireless communication network, a memory for storing various programs and data for encoding or decoding an image or inter or intra prediction for encoding or decoding, and executing a program And a microprocessor for controlling and the like. It can mean a variety of devices.
또한, 영상 부호화 장치에 의해 비트스트림으로 부호화된 영상은 실시간 또는 비실시간으로 인터넷, 근거리 무선 통신망, 무선랜망, 와이브로망, 이동통신망 등의 유무선 통신망 등을 통하거나 케이블, 범용 직렬 버스(USB: Universal Serial Bus) 등과 같은 다양한 통신 인터페이스를 통해 영상 복호화 장치로 전송되어 영상 복호화 장치에서 복호화되어 영상으로 복원되고 재생될 수 있다.In addition, the image encoded in the bitstream by the video encoding apparatus is real-time or non-real-time through the wired or wireless communication network, such as the Internet, local area wireless communication network, wireless LAN network, WiBro network, mobile communication network, or the like, or a cable, universal serial bus (USB: Universal) It may be transmitted to an image decoding apparatus through various communication interfaces such as a serial bus, and may be decoded by the image decoding apparatus to restore and reproduce the image.
통상적으로 동영상은 일련의 픽처(Picture)로 구성될 수 있으며, 각 픽처들은 프레임 또는 블록(Block)과 같은 소정의 영역으로 분할될 수 있다. 영상의 영역이 블록으로 분할되는 경우에는 분할된 블록은 부호화 방법에 따라 크게 인트라 블록(Intra Block), 인터 블록(Inter Block)으로 분류될 수 있다. 인트라 블록은 인트라 예측 부호화(Intra Prediction Coding) 방식을 사용하여 부호화되는 블록을 뜻하는데, 인트라 예측 부호화란 현재 부호화를 수행하는 현재 픽처 내에서 이전에 부호화되고 복호화되어 복원된 블록들의 화소를 이용하여 현재 블록의 화소를 예측함으로써 예측 블록을 생성하고 현재 블록의 화소와의 차분값을 부호화하는 방식이다. 인터 블록은 인터 예측 부호화(Inter Prediction Coding)를 사용하여 부호화되는 블록을 뜻하는데, 인터 예측 부호화란 하나 이상의 과거 픽처 또는 미래 픽처를 참조하여 현재 픽처 내의 현재 블록을 예측함으로써 예측 블록을 생성하고 현재 블록과의 차분값을 부호화하는 방식이다. 여기서, 현재 픽처를 부호화하거나 복호화하는데 참조되는 프레임을 참조 프레임(Reference Frame)이라고 한다.In general, a video may be composed of a series of pictures, and each picture may be divided into a predetermined area such as a frame or a block. When a region of an image is divided into blocks, the divided blocks may be classified into intra blocks and inter blocks according to an encoding method. An intra block refers to a block that is encoded by using an intra prediction coding scheme. Intra prediction coding is performed by using pixels of blocks that have been previously encoded, decoded, and reconstructed in a current picture that performs current encoding. A prediction block is generated by predicting pixels of a block, and a difference value with pixels of the current block is encoded. An inter block refers to a block that is encoded using inter prediction coding. Inter prediction coding generates a prediction block by predicting a current block in a current picture by referring to one or more past pictures or future pictures, and then generates a current block. This is a method of encoding the difference value with. Here, a frame referred to for encoding or decoding the current picture is referred to as a reference frame.
도 3은 휘도(Luma) 성분에서 통상적인 서브 샘플을 이용한 움직임 예측을 나타낸 도면이다.3 is a diagram illustrating motion prediction using a conventional subsample in a luminance (Luma) component.
도 3을 참조하면, 정수 샘플 단위에서 1/4 샘플의 서브 샘플 위치까지 움직임 벡터를 구할 수 있다. Referring to FIG. 3, a motion vector may be obtained from an integer sample unit up to a subsample position of 1/4 samples.
도 4는 6-tab FIR(Finite Impulse Response) 필터의 예시이다.4 is an illustration of a 6-tab Finite Impulse Response (FIR) filter.
도 4를 참조하면, 정수 화소의 값을 가지는 6개의 값과 6개의 필터계수 {1, -5, 20, 20, -5, 1} 를 이용하여 1/2 샘플 위치의 화소값을 보간하고 1/4 샘플의 성분은 보간된 1/2 샘플 화소값과 정수 화소값 또는 두개의 1/2 샘플 화소값을 이용한 선형 보간으로 보간하여 다음 식의 예시처럼 1/4 샘플의 위치의 화소값을 보간한다.Referring to FIG. 4, pixel values at 1/2 sample positions are interpolated using 6 values having integer pixel values and 6 filter coefficients {1, -5, 20, 20, -5, 1} and 1 The components of the / 4 sample are interpolated with interpolated 1/2 sample pixel values and integer pixel values or linear interpolation using two 1/2 sample pixel values to interpolate pixel values at positions of 1/4 samples as shown in the following equation. do.
예) b' = (1×E + (-5)×F + 20×G + 20×H + (-5)×I + 1×J)Ex) b '= (1 × E + (-5) × F + 20 × G + 20 × H + (-5) × I + 1 × J)
b = b' / 32b = b '/ 32
a = ((32×G) + b') / 64a = ((32 × G) + b ') / 64
이러한 서브 샘플들을 생성하는데 통상적인 이 방법 외에 더 정확한 서브 샘플의 값을 보간하기 위하여 1/2 샘플을 보간할 때 32로 나누지 않고 정수 화소의 값에 32를 곱하여 1/4 샘플의 값을 선형 보간하고 64로 나눠줌으로써 중간에 일어나는 나눗셈 연산으로 인한 오차 값을 줄이므로 더 정확한 서브샘플의 값을 보간하는 방법과 다른 필터 계수를 사용하는 등 많은 방법이 있다.In addition to this method of generating these subsamples, in order to interpolate the values of subsamples more precisely, when interpolating 1/2 samples, the integer pixel's value is multiplied by 32 instead of 32 and the value of 1/4 samples is linear interpolated. And dividing by 64 reduces the error value due to the division operation that occurs in the middle, so there are many methods such as interpolating more accurate subsample values and using different filter coefficients.
도 5는 색차(Chroma) 샘플에 대한 선형 보간을 예시한 도면이다.5 is a diagram illustrating linear interpolation for a chroma sample.
도 5를 참조하면, 다음 식의 예시처럼 정수 화소 4개의 값을 이용하여 양선형 보간(bi-linear interpolation)을 이용하여, 각 서브샘플의 위치를 고려하여 가중치 값을 곱해줌으로써 아래 수식처럼 1/8 서브 샘플의 화소값을 보간한다.Referring to FIG. 5, multiplying weight values by considering bi-linear interpolation using bi-linear interpolation using the values of four integer pixels as in the following example shows that 1 / The pixel values of the 8 subsamples are interpolated.
(수식): a = [(8-dx)×(8-dy)×A + dx×(8-dy)×B + (8-dx)×dy×C + dx×dy×D)]/64.(Formula): a = [(8-dx) × (8-dy) × A + dx × (8-dy) × B + (8-dx) × dy × C + dx × dy × D)] / 64 .
즉, 도 5의 예에서 {a = [(6×5×A)+(2×5×B)+(6×3×C)+(2×3×D)] / 64}가 된다.That is, in the example of FIG. 5, {a = [(6 × 5 × A) + (2 × 5 × B) + (6 × 3 × C) + (2 × 3 × D)] / 64}.
1/8 서브 샘플 보간을 하는 이유는 4:2:0 영상형태의 경우 색차신호의 크기가 휘도신호의 1/4 (가로 1/2, 세로 1/2)이기 때문이다.The reason why the 1/8 subsample interpolation is performed is that the color difference signal is 1/4 (horizontal 1/2, vertical 1/2) of the luminance signal in the 4: 2: 0 image form.
이처럼 색차 샘플에 대해서 선형 보간만을 사용하여 1/8 샘플의 해상도까지 보간하였을 경우 이는 6-tap FIR 필터와 비교해 보았을 때 정확하지 않기 때문에 색차 샘플을 부호화하는데 있어서 압축 효율을 저하시킨다.When the interpolation is performed using only linear interpolation for the color difference samples, the resolution of 1/8 samples is not accurate when compared to the 6-tap FIR filter, thereby reducing the compression efficiency in encoding the color difference samples.
도 6은 본 발명의 일 실시예에 따른 영상 부호화 장치를 간략하게 나타낸 블록 구성도이다.6 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.
본 발명의 일 실시예에 따른 영상 부호화 장치(600)는 영상의 현재 블록을 휘도 성분의 움직임 벡터 값을 이용하여 색차 성분의 서브샘플 예측값을 생성하여 영상을 부호화하는 장치로서, 도 6에 도시한 바와 같이, 예측부(610), 감산부(620), 변환부(630), 양자화부(640) 및 부호화부(650)를 포함하여 구성될 수 있다.The image encoding apparatus 600 according to an embodiment of the present invention encodes an image by generating a subsample prediction value of a chrominance component using a motion vector value of a luminance component of a current block of the image. As described above, the prediction unit 610 may include a predictor 610, a subtractor 620, a transformer 630, a quantizer 640, and an encoder 650.
부호화하고자 하는 입력 영상은 매크로 블록(Macro Block) 단위로 입력이 되는데, 본 발명에서, 매크로 블록은 M×N 형태이며 M과 N은 각 2n의 크기를 가지며, M과 N이 동일하거나 다를 수 있다.The input image to be encoded is input in units of macro blocks. In the present invention, the macro blocks are in the form of M × N, M and N each have a size of 2 n , and M and N may be the same or different. have.
예측부(610)는 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성한다.The prediction unit 610 generates a prediction block for the color difference component from the value obtained by interpolating a value of the subsample component of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block using an FIR filter and linear interpolation. do.
예측부(610)는 현재 매크로 블록을 예측하기 위하여 다른 프레임을 이용하여 예측 블록을 생성할 수 있다. 즉, 예측부(610)는 이미 부호화 과정을 거치고 복원된 이전 프레임에서 움직임 추정을 통해 움직임 벡터를 생성하고 이 움직임 벡터를 이용한 움직임 보상 과정에서 예측 블록을 생성할 수 있다. 이때, 예측부(610)는 휘도 성분의 움직임 벡터의 값을 이용하여 색차 성분에서도 동일한 움직임 벡터값을 사용할 수 있으며, 이 움직임 벡터가 지시하는 참조프레임 내의 블록의 서브샘플 성분의 값을 고정밀도 FIR 필터와 선형 보간을 사용한 예측 값을 갖는 예측 블록을 생성할 수 있다.The prediction unit 610 may generate a prediction block using another frame to predict the current macro block. That is, the prediction unit 610 may generate a motion vector through motion estimation in a previous frame that has already been encoded and reconstructed, and generate a prediction block in a motion compensation process using the motion vector. In this case, the prediction unit 610 may use the same motion vector value even in the color difference component by using the motion vector value of the luminance component, and use the high-precision FIR to determine the value of the subsample component of the block in the reference frame indicated by the motion vector. A prediction block having prediction values using a filter and linear interpolation may be generated.
감산부(620)는 현재 블록의 각 화소의 원 화소값과 예측부(610)가 생성한 예측값의 차이값을 계산하여 잔차신호를 생성한다.The subtractor 620 generates a residual signal by calculating a difference value between the original pixel value of each pixel of the current block and the predicted value generated by the predictor 610.
변환부(630)는 감산부(620)에 의해 생성된 잔차신호를 주파수 영역으로 변환한다. 여기서, 변환부(630)는 이산 코사인 변환(DCT: Discrete Cosine Transform, 이하 'DCT 변환'이라 칭함) 또는 웨이블릿 변환(Wavelet Transform) 등의 시간축의 화상 신호를 주파수축으로 변환하는 다양한 변환 기법을 이용하여 잔차신호를 주파수 영역으로 변환할 수 있다.The converter 630 converts the residual signal generated by the subtractor 620 into the frequency domain. Here, the transform unit 630 uses various transformation techniques for transforming an image signal of a time axis into a frequency axis, such as Discrete Cosine Transform (DCT) or Wavelet Transform. The residual signal can be converted into the frequency domain.
양자화부(640)는 변환부(630)에 의해 주파수 영역으로 변환된 잔차신호로 이루어진 주파수변환블록을 양자화(Quantization)한다. 양자화 방식으로는 데드존 균일 경계 양자화(DZUTQ: Dead Zone Uniform Threshold Quantization, 이하 'DZUTQ'라 칭함) 또는 양자화 가중치 매트릭스 (Quantization Weighted Matrix) 등과 같은 다양한 양자화 기법을 이용할 수 있다.The quantization unit 640 quantizes the frequency conversion block formed of the residual signal converted into the frequency domain by the conversion unit 630. As a quantization method, various quantization techniques such as dead zone uniform threshold quantization (DZUTQ) or quantization weighted matrix (Quantization Weighted Matrix) may be used.
부호화부(650)는 양자화부(640)에 의해 양자화된 주파수 계수로 이루어진 양자화된 주파수변환블록을 비트스트림으로 부호화한다. 이러한 부호화 기술로서는 엔트로피 부호화(Entropy Encoding) 기술이 사용될 수 있으나, 반드시 이에 한정하지 않고 다른 다양한 부호화 기술이 사용될 수도 있을 것이다.The encoder 650 encodes the quantized frequency transform block including the frequency coefficient quantized by the quantizer 640 into a bitstream. As the encoding technique, an entropy encoding technique may be used, but various encoding techniques may be used without being limited thereto.
또한, 부호화부(650)는 양자화 주파수 계수들을 부호화한 비트열뿐만 아니라 부호화된 비트열을 복호화하는 데 필요한 다양한 정보들을 부호화 데이터에 포함시킬 수 있다. 즉, 부호화 데이터는 부호화된 블록 형태(CBP: Coded Block Pattern), 델타 양자화 계수(Delta Quantization Parameter) 및 양자화 주파수 계수가 부호화 된 비트열이 포함되는 필드와 예측에 필요한 정보(예를 들어, 인트라 예측의 경우 인트라 예측 모드 또는 인터 예측의 경우 움직임 벡터 등)를 위한 비트가 포함되는 필드를 포함할 수 있다.In addition, the encoder 650 may include not only a bit string encoding the quantized frequency coefficients but also various pieces of information necessary to decode the encoded bit string in the encoded data. That is, the coded data includes a field including a coded block pattern (CBP), a delta quantization parameter, and a bit string in which the quantization frequency coefficients are encoded, and information necessary for prediction (eg, intra prediction). In the case of Intra prediction mode or a motion vector in the case of inter prediction, etc.) may include a field that contains a bit.
역양자화부(660)는 변환 및 양자화된 잔차 블록(즉, 양자화된 주파수변환블록)을 역 양자화(Inverse Quantization)하고, 역변환부(670)는 역양자화된 변환 잔차 블록을 역 변환(Inverse Transform)하여 잔차 블록을 복원(Reconstruction)한다. 여기서, 역 양자화와 역 변환은 변환부(630)가 수행한 변환 과정과 양자화부(640)가 수행한 양자화 과정을 각각 역으로 수행함으로써 이루어질 수 있다. 즉, 역양자화부(660) 및 역변환부(670)는 변환부(630) 및 양자화부(640)로부터 발생되어 전달되는 변환 및 양자화에 관한 정보(예를 들어, 변환 및 양자화 타입에 대한 정보)를 이용하여 역 양자화 및 역 변환을 수행할 수 있다.The inverse quantization unit 660 inverse quantizes the transformed and quantized residual block (ie, the quantized frequency transform block), and the inverse transform unit 670 inverses the inverse quantized transform residual block. To reconstruct the residual block. Here, inverse quantization and inverse transformation may be performed by inversely performing a transformation process performed by the transform unit 630 and a quantization process performed by the quantization unit 640. That is, the inverse quantization unit 660 and the inverse transform unit 670 are information on transform and quantization generated and transmitted from the transform unit 630 and the quantization unit 640 (for example, information on a transform and quantization type). Inverse quantization and inverse transformation may be performed using.
가산부(680)는 예측부(610)에서 생성된 예측 블록과 역변환부(670)에 의해 생성된 잔차 블록을 가산하여 복원된 블록을 생성한다.The adder 680 adds the prediction block generated by the predictor 610 and the residual block generated by the inverse transform unit 670 to generate a reconstructed block.
프레임 메모리(690)는 가산부(680)에서 복원된 블록을 저장하여 인트라 혹은 인터 예측 수행 시 예측 블록을 생성하기 위해 참조블록으로 사용된다.The frame memory 690 stores a block reconstructed by the adder 680 and is used as a reference block to generate a prediction block when performing intra or inter prediction.
도 7은 본 발명의 일 실시예에 따라 예측부(610)가 블록 내의 서브샘플값을 보간하는 과정을 설명하기 위한 예시도이다.7 is an exemplary diagram for describing a process of the prediction unit 610 interpolating a subsample value in a block according to an embodiment of the present invention.
도 7의 서브샘플의 보간은 수학식 1 내지 수학식 4와 같은 방법을 이용하여 수행될 수 있다.Interpolation of the subsample of FIG. 7 may be performed using a method such as Equation 1 to Equation 4.
수학식 1
Figure PCTKR2011007418-appb-M000001
 Equation 1
Figure PCTKR2011007418-appb-M000001
즉, 1/2 서브샘플 값은 주변의 정수화소값에 대하여 각각 소정의 값만큼 곱한 후 이들 곱해진 값들을 더한 값인 1/2샘플 증폭값을 이용하여 구할 수 있다.That is, the 1/2 subsample value can be obtained by multiplying the integer pixel values by a predetermined value and then using the 1/2 sample amplification value which is the sum of these multiplied values.
즉, 1/2 서브샘플 값은 좌측 3개의 정수화소 값(A, B, C)과 우측 3개의 정수화소 값(D, E, F)을 이용하여 구할 수 있다.That is, the 1/2 subsample value can be obtained by using three left integer pixel values A, B, and C and three right integer pixel values D, E, and F.
수학식 2
Figure PCTKR2011007418-appb-M000002
 Equation 2
Figure PCTKR2011007418-appb-M000002
즉, 1/4 서브샘플 값은 최근접 정수화소 값(C)과 1/2샘플 증폭값(d')을 이용하여 구할 수 있다. b를 수식의 구하는 형식은 1/2 서브샘플을 직접 선형보간하지 않고 최근접 정수화소 값(C)과 1/2샘플 증폭값(d')을 선형보간한 형태를 다시 나누어서 d를 구하면서 발생하는 정수화에 의한 오차를 제거할 수 있다.That is, the 1/4 subsample value can be obtained using the nearest integer pixel value C and the half sample amplification value d '. The formula for obtaining b is obtained by dividing the form of linear interpolation between the nearest integer pixel value (C) and the half-sample amplification value (d ') instead of directly linearly interpolating the 1/2 subsample. The error due to the integerization can be eliminated.
수학식 3
Figure PCTKR2011007418-appb-M000003
 Equation 3
Figure PCTKR2011007418-appb-M000003
b' 역시 최근접 정수화소 값(C)과 1/2샘플 증폭값(d')을 이용하여 구할 수 있으므로, 1/8 서브샘플 값 역시 최근접 정수화소 값(C)과 1/2샘플 증폭값(d')을 이용하여 구할 수 있다.b 'can also be obtained using the nearest integer pixel value (C) and the half-sample amplification value (d'), so that the 1/8 subsample value is also the nearest integer pixel value (C) and half-sample amplification. Can be obtained using the value d '.
수학식 4
Figure PCTKR2011007418-appb-M000004
 Equation 4
Figure PCTKR2011007418-appb-M000004
수학식 1, 2, 3, 4 및 도 7에서, A, B, C, D, E, F는 색차 성분의 정수 화소 값을 나타내며 d는 색차성분의 1/2 서브샘플, b는 색차성분의 1/4 서브샘플, a는 색차성분의 1/8 서브샘플을 나타낸다.In Equations 1, 2, 3, 4, and 7, A, B, C, D, E, and F represent integer pixel values of the color difference component, d represents 1/2 subsample of the color difference component, and b represents the color difference component. 1/4 subsample, a, represents 1/8 subsample of the color difference component.
수학식 1에 나타낸 바와 같이, 색차성분의 1/2 서브샘플의 값은 고정밀도 필터(여기서 고정밀도 필터는 FIR필터 등 다양한 필터를 사용할 수 있음)를 이용하여 생성하고, 수학식 2에 나타낸 바와 같이, 1/4 서브샘플의 값은 고정밀도 필터 및 선형보간을 사용한다. 특히 1/8 서브샘플의 경우에는 수학식 3과 같이 정수 화소와 1/4 서브샘플을 이용하여 고정밀도의 선형 보간할 수 있으며, 또한 수학식 4와 같이 정수화소와 1/2 서브샘플을 이용하여 고정밀도의 선형 보간을 할 수 있다. 이때, w1과 w2는 정수화소와 1/2 서브샘플에 곱해지는 가중치 값을 나타낸다. 이와 같이 수학식 1을 이용한 1/2 서브샘플 값(예컨대 d)과, 수학식 2를 이용한 1/4 서브샘플 값(예컨대 b)은 고정밀도 FIR필터를 사용하여 생성하고, 수학식 3을 이용한 1/8 서브샘플 값(예컨대 a)은 두 화소값 사이의 선형 보간을 통해 생성하므로 모든 서브샘플을 선형 보간만으로 보간하였을 경우보다 더 정확한 값을 생성할 수 있다.As shown in Equation 1, the value of the 1/2 subsample of the chrominance component is generated by using a high precision filter (where the high precision filter can use various filters such as FIR filter), and as shown in Equation 2 Similarly, the values of quarter subsamples use high precision filters and linear interpolation. In particular, in the case of 1/8 subsample, high-precision linear interpolation can be performed using integer pixels and 1/4 subsamples as shown in Equation 3, and also integer pixels and 1/2 subsamples are used as shown in Equation 4. High accuracy linear interpolation can be achieved. At this time, w1 and w2 represent weight values multiplied by integer pixels and 1/2 subsample. As described above, the 1/2 subsample value (e.g., d) using Equation 1 and the 1/4 subsample value (e.g., b) using Equation 2 are generated using a high-precision FIR filter, and Equation 3 is used. Since a 1/8 subsample value (eg, a) is generated through linear interpolation between two pixel values, all subsample values can be generated more accurately than when interpolation is performed using only linear interpolation.
수학식 1 내지 5에서 사용되는 나눗셈 연산은 모두 비트 시프트 연산 (>>, <<)을 이용하여 처리 속도를 향상시킬 수 있다.All of the division operations used in Equations 1 to 5 can improve the processing speed by using bit shift operations (>>, <<).
또한, 수학식 1 내지 4에서 소수점 이하에 대한 반올림 연산을 위하여 제수(나누는 값)의 절반을 피제수에 미리 더하여 사용될 수 있다. 다음 수학식 5는 수학식 3에 상기 반올림 연산을 추가한 식이다.In addition, half of the divisor (dividing value) may be used in addition to the dividend in advance for the rounding operation to below the decimal point in Equations 1 to 4. Equation 5 is an equation in which the above rounding operation is added to Equation 3 below.
수학식 5
Figure PCTKR2011007418-appb-M000005
 Equation 5
Figure PCTKR2011007418-appb-M000005
여기서, FIR필터는 디지털 필터의 한 종류로 입력신호의 일정한 값들만을 가지고 필터링을 수행한다. 따라서 FIR필터에 대하여 필터의 특성함수인 임펄스 응답을 구해보면 유한한 길이를 가진다. 또한 FIR필터의 식의 형태에서 보면 회귀(feedback)성분을 갖지 않으므로 동일한 특성을 구현할 때 차수가 높아져서 실행시간이 증가하지만 위상변이(즉, 입력과 출력간의 파형의 형태유지)가 중요한 경우에는 FIR필터를 사용할 수 있다. Here, the FIR filter is a type of digital filter and performs filtering with only certain values of the input signal. Therefore, for the FIR filter, the impulse response, which is the characteristic function of the filter, has a finite length. In addition, in the form of the FIR filter, it does not have a feedback component, so the execution time increases due to the higher order when implementing the same characteristic, but the FIR filter is important when phase shift (ie, the shape maintenance of the waveform between the input and the output) is important. Can be used.
또한, 여기서 사용한 고정밀도 FIR필터는 1/2 서브샘플은 FIR필터를 사용하고, 1/4 서브샘플은 수학식 2에서 나타낸 것과 같이 FIR필터를 사용한 값(d')에 32로 나눈 1/2 서브샘플의 값(수학식 1에서의 d)을 사용하지 않고 32로 나누기 전의 값(즉, 1/2 서브샘플의 값을 구할때 얻은 FIR필터링 결과값 d')과 정수화소의 값에 32를 곱한 값(32XC)을 선형 보간함으로써 중간에 나눗셈 연산으로 인한 정보손실을 막음으로써 더 정밀한 값을 보간하는 방법이다.In addition, the high precision FIR filter used here is a 1/2 subsample using an FIR filter, and a 1/4 subsample is 1/2 divided by 32 to the value (d ') using the FIR filter as shown in Equation 2. Rather than using the subsample value (d in Equation 1), multiply the value before dividing by 32 (that is, the FIR filtering result d 'obtained by obtaining the value of 1/2 subsample) and the integer pixel value by 32. Linear interpolation of the value 32XC prevents the loss of information due to the division operation in the middle and thus interpolates more precise values.
또한 수학식 1의 d', 수학식 2의 b', 수학식 3의 a' 와 같은 모든 서브샘플값들과, 서브샘플값과 같은 비트수를 갖도록 비트를 증가시킨 정수 화소값은 다음단계에서 현재블록에 대한 예측블록을 얻을 때 샘플 값들로 사용할 수 있으며, 예측블록을 얻은 후 가중치로 나누어 예측블록의 샘플당 비트수가 현재블록의 샘플당 비트수가 되도록 함으로서 더 좋은 성능을 보이는 것이 가능하다.In addition, all subsample values such as d 'in Equation 1, b' in Equation 2, and a 'in Equation 3, and integer pixel values in which bits are increased to have the same number of bits as the subsample value are described in the next step. When the prediction block for the current block is obtained, the prediction block may be used as sample values. By obtaining the prediction block and dividing by the weight, the number of bits per sample of the prediction block may be better than the number of bits per sample of the current block.
또한 수학식 4에서 a(1/8 서브샘플)를 보간한 값만을 구하는 식인 것으로 설명하였으나 b(1/4 서브샘플)을 구하는 경우에도 a를 구하는 경우와 달리하여 가중치(w1, w2)를 설정함으로써 b를 구할 수 있다. 그리고, 마찬가지의 방법으로 수학식 2에서 b' 대신 f', b 대신 f, C 대신 D를 사용하면 f를 구할 수 있으며, 수학식 3에서 a' 대신 g, a 대신 g, b' 대신 f', b 대신 f, C 대신 D를 사용하면 g를 구할 수 있다. 한편, 1/8 서브샘플인 c는 b와 d를 보간하면 구할 수 있고, e는 d와 f를 보간하면 구할 수 있으므로 a 또는 g와 유사하게 최근접 정수화소 값(C)과 1/2샘플 증폭값(d')을 이용하여 구할 수 있다.In addition, in Equation 4, only the value obtained by interpolating a (1/8 subsample) is described. However, the weight (w1, w2) is set differently from the case of obtaining a even when obtaining b (1/4 subsample). B can be obtained. In the same manner, f can be obtained by using f 'instead of b', f instead of b, and D instead of C in Equation 2, and g 'instead of a', g 'instead of a, and f' instead of b 'in Equation 3 You can get g by using f instead of b and D instead of C. On the other hand, 1/8 subsample c can be obtained by interpolating b and d, and e can be obtained by interpolating d and f, so that the nearest integer pixel value (C) and 1/2 sample are similar to a or g. It can be obtained using the amplification value d '.
위와 같이 주변의 정수화소 값을 구하여 1/2 서브샘플을 구하고, 이보다 더 세밀한 서브샘플의 값(1/4 서브샘플, 1/8 서브샘플 등)은 좌우의(또는 상하의) 덜 세밀한 근접 서브샘플 값 또는 근접 정수화소 값을 이용하여 보간(즉, 1/4 샘플값을 구할 때는 1/2서브샘플 및 정수화소를 이용)할 수 있다. 이와 같이 다양한 위치의 서브샘플의 보간을 위하여 서브샘플 또는 정수화소의 값을 사용하는 경우에 대한 것은 모든 서브샘플의 경우에 대하여 설명하지 않더라도 당업자라면 이를 추론할 수 있을 것이다.As shown above, the surrounding integer pixel value is obtained to obtain 1/2 subsample, and the value of the finer subsample (1/4 subsample, 1/8 subsample, etc.) is smaller than the left and right (or the top and bottom) subtle subsamples. Interpolation (i.e., using 1/2 subsample and integer pixel to obtain a 1/4 sample value) can be performed using the value or the adjacent integer pixel value. As described above, the case of using the value of the subsample or the integer pixel for interpolation of the subsamples of various positions may be inferred by those skilled in the art even if the case of all the subsamples is not described.
또한, 여기서 사용한 고정밀도 FIR필터와 선형 보간 이외에 다른 보간법을 사용하여도 되지만, 1/2 샘플을 위하여 6-tab 이상을 갖는 FIR 필터, 1/4 샘플을 위하여 위의 수학식 2처럼 고정밀도 기능을 갖도록 보간을 하는 것이 중요하다.In addition to the high-precision FIR filter and linear interpolation used here, other interpolation methods may be used, but FIR filters having 6-tab or more for 1/2 samples, and high-precision functions such as Equation 2 above for 1/4 samples It is important to interpolate to have.
4:2:0의 영상형태의 경우 위의 수학식 1 내지 수학식 3을 이용하여 참조프레임 내의 참조블록을 보간할 수 있고, 4:4:4의 영상형태의 경우는 참조블록 내의 휘도 성분과 색차 성분의 크기가 동일하므로 휘도 성분과 마찬가지로 색차 성분 또한 1/4샘플 위치까지 보간하면 된다. 그러므로 수학식 3의 1/8 샘플위치의 보간은 생략하면 4:4:4 영상형태에서도 사용 가능하다.In the case of the 4: 2: 0 picture form, the reference blocks in the reference frame may be interpolated using Equations 1 to 3, and in the case of the 4: 4: 4 picture form, the luminance component in the reference block Since the magnitudes of the chrominance components are the same, the chrominance components, like the luminance component, may be interpolated to a quarter sample position. Therefore, if the 1/8 sample position interpolation of Equation 3 is omitted, it can be used in 4: 4: 4 image form.
또한, 4:2:2의 영상형태일 경우는 수평방향 색차 성분의 크기가 휘도 성분의 1/2이므로 수평방향의 색차 성분은 수학식 1 내지 수학식 3처럼 1/8 서브샘플 위치까지 보간하고, 수직 방향은 위에 4:4:4 영상형태와 마찬가지로 수학식 1 및 2를 이용하여 1/4 서브샘플의 위치까지만 보간하여 서브샘플을 생성할 수 있다.In the case of the 4: 2: 2 image, since the magnitude of the horizontal color difference component is 1/2 of the luminance component, the horizontal color difference component is interpolated to the 1/8 subsample position as shown in Equations 1 to 3 below. In the vertical direction, the subsample may be generated by interpolating up to the position of the 1/4 subsample using Equations 1 and 2 similarly to the 4: 4: 4 image form.
또한, 현재 HD급 이상의 화질에 대한 표준화 회의중인 TMuC(Test Model under Consideration)에서는 휘도 성분을 1/8 서브샘플까지 보간할 수 있으므로 4:2:0 영상형태에서 색차 성분은 1/16 서브샘플까지 보간할 수 있다. 그렇기 때문에 1/8 서브위치까지는 본 발명의 일 실시예에서 진술한 바와 같이 만들어 내고 1/16 서브샘플의 위치는 다시 선형 보간을 사용하여 보간할 수 있다.In addition, TMuC (Test Model under Consideration), which is currently in the standardization meeting for HD quality or higher, can interpolate luminance components to 1/8 subsamples, so that the chrominance components can be up to 1/16 subsamples in 4: 2: 0 video format. You can interpolate. Thus, up to 1/8 sub-positions can be made as stated in one embodiment of the present invention and the positions of 1/16 subsamples can be interpolated again using linear interpolation.
한편, 본 실시예에서 색차성분에 대한 것을 예로 들어 필터링 및 보간방법에 대하여 설명하였으나, 이러한 방법은 색차성분 뿐만 아니라 휘도성분 및 R, G, B 색상의 블록 등 다양한 블록에 대해서도 적용될 수 있는 것이며, 이는 후술하는 복호화방법에서 있어서도 마찬가로 색차성분 뿐만 아니라 휘도성분 및 R, G, B 색상의 블록 등 다양한 블록에 대해서도 적용될 수 있는 것이다.Meanwhile, in the present embodiment, the filtering and interpolation methods have been described with reference to color difference components as an example. However, the method may be applied to various blocks such as luminance components and blocks of R, G, and B colors as well as color difference components. The same can be applied to various blocks such as luminance components and blocks of R, G, and B colors as well as color difference components in the decoding method described later.
도 8은 본 발명의 일 실시예에 따른 영상 복호화 장치의 구성을 나타낸 블록 구성도이다.8 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.
도 8에 도시한 바와 같이, 본 발명의 일 실시예에 따른 영상 복호화 장치(800)는 영상의 현재 블록을 휘도 성분의 움직임 벡터 값을 이용하여 색차 성분의 서브샘플 예측값을 생성하여 복호화하는 장치로서, 복호화부(810), 역양자화부(820), 역변환부(830), 가산부(840) 및 예측부(850)를 포함하여 구성된다.As shown in FIG. 8, the image decoding apparatus 800 according to an embodiment of the present invention is an apparatus for generating and decoding a subsample prediction value of a chrominance component using a motion vector value of a luminance component of a current block of an image. And a decoder 810, an inverse quantizer 820, an inverse transformer 830, an adder 840, and a predictor 850.
복호화부(810)는 비트스트림을 복호화하여 양자화된 주파수변환블록을 추출한다.The decoder 810 extracts a quantized frequency transform block by decoding the bitstream.
복호화부(810)는 부호화 데이터를 복호화하여 양자화된 주파수변환블록뿐만 아니라 복호화에 필요한 정보들을 복호화하거나 추출할 수 있다. 복호화에 필요한 정보들은 부호화 데이터(즉, 비트스트림) 내의 부호화된 비트열을 복호화하는 데 필요한 정보들을 말하며, 예를 들어 블록 타입에 대한 정보, 움직임 벡터에 대한 정보, 변환 및 양자화 타입에 대한 정보 등이 될 수 있으며, 이외의 다양한 정보들이 될 수 있다.The decoder 810 may decode and extract not only the quantized frequency transform block but also information necessary for decoding by decoding the encoded data. Information necessary for decoding refers to information necessary for decoding the coded bit string in the encoded data (ie, the bitstream). For example, information about a block type, information about a motion vector, information about a transform and quantization type, and the like. It can be, and a variety of other information can be.
즉, 복호화부(810)는 영상 부호화 장치(600)에 의해 부호화된 데이터인 비트스트림을 복호화하여 영상의 현재 블록에 대한 화소 정보를 포함하고 있는 양자화된 주파수변환 블록을 추출하고, 추출된 예측에 필요한 정보를 예측부(850)로 전달한다.That is, the decoder 810 decodes a bitstream, which is data encoded by the image encoding apparatus 600, extracts a quantized frequency transform block including pixel information of the current block of the image, and extracts the extracted prediction signal. The necessary information is transmitted to the prediction unit 850.
예측부(850)는 복호화부(810)로부터 전달된 예측에 필요한 정보를 이용하여 영상 부호화 장치(600)의 예측부(610)와 동일한 방식으로 현재 블록을 예측할 수 있다.The prediction unit 850 may predict the current block in the same manner as the prediction unit 610 of the image encoding apparatus 600 by using the information necessary for the prediction transmitted from the decoder 810.
예측부(850)는 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성한다. 여기서 현재 블록의 색차 성분의 움직임 벡터를 구함에 있어서, 비트스트림에서 복원된 휘도 성분의 움직임 벡터 값을 이용하며, 이 움직임 벡터에 의하여 참조되는 참조블록의 색차 성분의 정수 화소값을 고정밀도 FIR필터와 선형 보간을 사용하여 예측값을 생성한다.The prediction unit 850 generates a prediction block for the color difference component from the value obtained by interpolating the values of the subsample components of the color difference component of the reference block referred to by the motion vector of the color difference component of the current block using an FIR filter and linear interpolation. do. In obtaining a motion vector of the chrominance component of the current block, the motion vector value of the luminance component reconstructed in the bitstream is used, and the integer pixel value of the chrominance component of the reference block referred to by the motion vector is converted into a high precision FIR filter. Generate predictions using and linear interpolation.
본 발명의 일 실시예에 따른 영상 복호화 장치(800)의 예측부(850)는 도 6을 통해 전술한 영상 부호화 장치(600)의 예측부(610)와 동일하게 서브샘플을 생성한다. 따라서, 중복 설명을 피하기 위해 이에 대한 상세한 설명은 생략한다.The prediction unit 850 of the image decoding apparatus 800 according to an embodiment of the present invention generates a subsample in the same manner as the prediction unit 610 of the image encoding apparatus 600 described above with reference to FIG. 6. Therefore, detailed description thereof will be omitted to avoid redundant description.
역양자화부(820)는 복호화부(810)에 의해 비트스트림으로부터 추출된 양자화된 주파수변환 블록을 역 양자화한다. 역변환부(830)는 역양자화부(820)에 의해 역 양자화된 주파수변환 블록을 시간 영역으로 역 변환한다.The inverse quantizer 820 inverse quantizes the quantized frequency transform block extracted from the bitstream by the decoder 810. The inverse transformer 830 inversely transforms the frequency transform block inversely quantized by the inverse quantizer 820 into the time domain.
가산부(840)는 역변환부(830)에 의해 역 변환되어 복원된 잔차신호와 예측부(850)에 의해 예측 화소값을 더하여 현재 블록의 원 화소값을 복원한다. 가산부(840)에 의해 복원된 현재 블록은 프레임 메모리(860)로 전달되어, 예측부(850)에서 다른 블록을 예측하는 데 활용될 수 있다.The adder 840 reconstructs the original pixel value of the current block by adding the residual pixel restored by the inverse transform by the inverse transformer 830 and the predicted pixel value by the predictor 850. The current block reconstructed by the adder 840 may be transferred to the frame memory 860 and used by the predictor 850 to predict another block.
프레임 메모리(860)는 복원된 영상을 저장하여 인트라 예측 블록 및 인터 예측 블록 생성을 가능하게 한다.The frame memory 860 stores the reconstructed image to enable generation of the intra prediction block and the inter prediction block.
본 발명의 일 실시예에 따른 영상 부호화/복호화 장치는 도 6의 영상 부호화 장치(600)의 비트스트림 출력단을 도 8의 영상 복호화 장치(800)의 비트스트림 입력단에 연결하여 구성될 수 있다.An image encoding / decoding apparatus according to an embodiment of the present invention may be configured by connecting a bitstream output terminal of the image encoding apparatus 600 of FIG. 6 to a bitstream input terminal of the image decoding apparatus 800 of FIG. 8.
본 발명의 일 실시예에 따른 영상 부호화/복호화 장치는, 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화기 및 비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록의 색차성분을 복원하는 영상 복호화기를 포함한다.In an image encoding / decoding apparatus according to an embodiment of the present invention, a value obtained by interpolating a value of a subsample component of a chrominance component of a reference block referred to by a motion vector of a chrominance component of a current block using an FIR filter and linear interpolation An image of generating a prediction block for a chrominance component from the chrominance component, subtracting the prediction block from the chrominance component of the current block, generating a residual block, transforming and quantizing the residual block, and generating a quantized frequency transform block and encoding the bitstream A quantized frequency transform block is generated from an encoder and a bitstream, and the quantized frequency transform block is inversely quantized and inverse transformed so that the color difference component of the reference block referred to by the motion vector of the color difference component of the current block to reconstruct and reconstruct the residual block. Interpolated values of subsample components using FIR filter and linear interpolation And a video decoder for generating a prediction block for the chrominance component from and reconstructing the chrominance component of the current block to be reconstructed by adding the reconstructed residual block and the generated prediction block.
여기서 영상 부호화기는 본 발명의 일 실시예에 따른 영상 부호화 장치(600)로 구현 가능하며, 영상 복호화기는 본 발명의 일 실시예에 따른 영상 복호화 장치(800)로 구현 가능하다.The image encoder may be implemented by the image encoding apparatus 600 according to an embodiment of the present invention, and the image decoder may be implemented by the image decoding apparatus 800 according to the embodiment of the present invention.
도 9는 본 발명의 일 실시예에 따른 영상 부호화 방법을 설명하기 위한 순서도이다.9 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.
즉, 영상 부호화 장치(600)는 영상의 현재 블록을 휘도 성분의 움직임 벡터 값을 이용하여 색차 성분의 서브샘플 예측값을 생성하는 예측단계(S910), 현재 블록의 원 화소값과 예측 화소값의 차이값을 계산하여 잔차신호를 생성하는 감산단계(S920), 생성된 잔차신호를 DCT 변환 또는 웨이블릿 변환 등을 이용하여 주파수 영역으로 변환하는 변환단계(S930), 주파수 영역으로 변환된 잔차신호를 양자화하는 양자화단계(S940), 양자화된 주파수변환 잔차신호를 비트스트림으로 부호화하는 부호화단계(S950)를 거쳐 영상을 부호화한다.That is, the image encoding apparatus 600 generates a subsample prediction value of the color difference component using the motion vector value of the luminance component of the current block of the image (S910), and the difference between the original pixel value and the prediction pixel value of the current block. A subtraction step (S920) of generating a residual signal by calculating a value, a conversion step (S930) of converting the generated residual signal into a frequency domain using a DCT transform or a wavelet transform, and quantizing the residual signal transformed into a frequency domain The image is encoded through a quantization step S940 and an encoding step S950 of encoding the quantized frequency transform residual signal into a bitstream.
여기서, 예측단계(S910)는 예측부(610)의 기능에, 감산단계(S920)는 감산부(620)의 기능에, 변환단계(S930)는 변환부(630)의 기능에, 양자화단계(S940)는 양자화부(640)의 기능에, 부호화단계(S950)는 부호화부(650)의 기능에 각각 대응되므로 상세한 설명은 생략한다.Here, the prediction step (S910) is a function of the prediction unit 610, the subtraction step (S920) is a function of the subtraction unit 620, the transform step (S930) is a function of the transformer 630, the quantization step ( Since S940 corresponds to the function of the quantization unit 640 and the encoding step S950 corresponds to the function of the encoder 650, detailed description thereof will be omitted.
도 10은 본 발명의 일 실시예에 따른 영상 복호화 방법을 설명하기 위한 순서도이다.10 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.
유무선 통신망 또는 케이블 등을 통해 영상에 대한 비트스트림을 수신하여 저장한 영상 복호화 장치(800)는 사용자의 선택 또는 실행 중인 다른 프로그램의 알고리즘에 따라 영상을 재생하기 위해, 영상의 현재 블록을 휘도 성분의 움직임 벡터 값을 이용하여 색차 성분의 서브샘플 예측값을 생성하고 영상을 복호화하여 복원한다.The video decoding apparatus 800, which receives and stores a bitstream of an image through a wired or wireless communication network or a cable, stores the current block of the image in order to reproduce the image according to a user's selection or an algorithm of another program being executed. A subsample prediction value of the chrominance component is generated using the motion vector value, and the image is decoded and reconstructed.
이를 위해, 영상 복호화 장치(800)는 비트스트림을 복호화하여 영상의 현재 블록의 화소값에 대한 정보를 나타내는 양자화된 주파수변환 잔차신호를 추출하는 복호화단계(S1010). 양자화된 주파수변환 잔차신호를 역 양자화하는 역양자화단계(S1020), 역 양자화된 주파수변환 잔차신호를 시간 영역으로 역변환하는 역변환단계(S1030), 시간 영역으로 역 변환되어 복원된 잔차신호가 나타내는 현재 블록의 예측값을 휘도 성분의 움직임 벡터 값을 이용하여 색차 성분의 서브샘플 예측값을 생성하는 예측단계(S1040), 단계 S1030에서 복원된 현재 블록의 잔차신호와 단계 S1040에서 예측된 현재 블록의 각 화소의 예측 화소값을 더하여 현재 블록의 원 화소값을 복원하는 가산단계(S1050)를 거쳐 전송된 비트스트림을 복호화한다.To this end, the image decoding apparatus 800 decodes the bitstream and extracts a quantized frequency transform residual signal representing information about the pixel value of the current block of the image (S1010). An inverse quantization step of inversely quantizing a quantized frequency transform residual signal, an inverse transform step of inversely transforming an inverse quantized frequency transform residual signal into a time domain, and a current block represented by a residual signal inversely transformed into a time domain and restored A prediction step (S1040) of generating a subsample prediction value of a chrominance component using the motion vector value of the luminance component of the predicted value of, and prediction of each pixel of the current block predicted in step S1040 and the residual signal of the current block restored in step S1040. The bitstream transmitted through the addition step S1050 of restoring the original pixel value of the current block by adding the pixel value is decoded.
여기서, 복호화단계(S1010)는 복호화부(810)의 동작에 대응되며, 역양자화단계(S1020)는 역양자화부(820)의 동작에 대응되며, 역변환단계(S1030)는 역변환부(830)의 동작에 대응되며, 예측단계(S1040)는 예측부(850)의 동작에 대응되며, 가산단계(S1050)는 가산부(840)의 동작에 대응되므로 상세한 설명은 생략한다.Here, the decoding step (S1010) corresponds to the operation of the decoding unit 810, the inverse quantization step (S1020) corresponds to the operation of the inverse quantization unit 820, and the inverse transform step (S1030) of the inverse transformer (830). Corresponding to the operation, the prediction step (S1040) corresponds to the operation of the prediction unit 850, and the addition step (S1050) corresponds to the operation of the adding unit 840, and thus, detailed description thereof will be omitted.
본 발명의 일 실시예에 따른 영상 부호화/복호화 방법은, 본 발명의 일 실시예에 따른 영상 부호화 방법과 본 발명의 일 실시예에 따른 영상 복호화 방법을 결합하여 구현함으로써 실현할 수 있다.An image encoding / decoding method according to an embodiment of the present invention may be realized by combining the image encoding method according to an embodiment of the present invention and the image decoding method according to an embodiment of the present invention.
본 발명의 일 실시예에 따른 영상 부호화/복호화 방법은, 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 현재 블록의 색차 성분에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화단계 및 비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 색차 성분의 움직임 벡터에 의해 참조되는 참조블록의 색차성분의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 색차 성분에 대한 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록의 색차성분을 복원하는 영상 복호화단계를 포함한다.In an image encoding / decoding method according to an embodiment of the present invention, a value obtained by interpolating a value of a subsample component of a chrominance component of a reference block referred to by a motion vector of a chrominance component of a current block using an FIR filter and linear interpolation An image of generating a prediction block for a chrominance component from the chrominance component, subtracting the prediction block from the chrominance component of the current block, generating a residual block, transforming and quantizing the residual block, and generating a quantized frequency transform block and encoding the bitstream A color difference component of a reference block referred to by a motion vector of a color difference component of a current block to generate and quantize the quantized frequency transform block from the encoding step and the bitstream, and to inversely quantize and inverse transform the quantized frequency transform block. The subsample components of are interpolated using FIR filter and linear interpolation. And generating a prediction block for the color difference component from the value, and adding the reconstructed residual block and the generated prediction block to reconstruct the color difference component of the current block to be reconstructed.
여기서 영상 부호화단계는 본 발명의 일 실시예에 따른 영상 부호화 단계로 구현 가능하며, 영상 복호화단계는 본 발명의 일 실시예에 따른 영상 복호화 단계로 구현 가능하다.The image encoding step may be implemented as an image encoding step according to an embodiment of the present invention, and the image decoding step may be implemented as an image decoding step according to an embodiment of the present invention.
본 발명의 실시예를 구성하는 모든 구성 요소들이 각각 하나의 독립적인 하드웨어로 구현될 수 있지만, 각 구성 요소들의 그 일부 또는 전부가 선택적으로 조합되어 하나 또는 복수 개의 하드웨어에서 조합된 일부 또는 전부의 기능을 수행하는 프로그램 모듈을 갖는 컴퓨터 프로그램으로서 구현될 수도 있다. 그 컴퓨터 프로그램을 구성하는 코드들 및 코드 세그먼트들은 본 발명의 기술 분야의 당업자에 의해 용이하게 추론될 수 있을 것이다. 이러한 컴퓨터 프로그램은 컴퓨터가 읽을 수 있는 저장매체(Computer Readable Media)에 저장되어 컴퓨터에 의하여 읽혀지고 실행됨으로써, 본 발명의 실시예를 구현할 수 있다. 컴퓨터 프로그램의 저장매체로서는 자기 기록매체, 광 기록매체, 캐리어 웨이브 매체 등이 포함될 수 있다.Although all the components constituting the embodiment of the present invention may be implemented in one independent hardware, each of some or all of the components are selectively combined to some or all of the functions combined in one or a plurality of hardware. It may be implemented as a computer program having a program module for performing the operation. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.
이상의 설명은 본 발명의 기술 사상을 예시적으로 설명한 것에 불과한 것으로서, 본 발명이 속하는 기술 분야에서 통상의 지식을 가진 자라면 본 발명의 본질적인 특성에서 벗어나지 않는 범위에서 다양한 수정 및 변형이 가능할 것이다. 따라서, 본 발명에 개시된 실시예들은 본 발명의 기술 사상을 한정하기 위한 것이 아니라 설명하기 위한 것이고, 이러한 실시예에 의하여 본 발명의 기술 사상의 범위가 한정되는 것은 아니다. 본 발명의 보호 범위는 아래의 청구범위에 의하여 해석되어야 하며, 그와 동등한 범위 내에 있는 모든 기술 사상은 본 발명의 권리범위에 포함되는 것으로 해석되어야 할 것이다.The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of the present invention.
이상에서 설명한 바와 같이 본 발명의 실시예에 의하면, 영상의 서브샘플 예측값을 생성하는데 있어서, 실제의 성분과 예측된 성분 간의 차이를 최소화하는 보간 및 그를 이용한 영상의 부호화/복호화 분야에 적용되어, 부호화하고자 하는 현재 블록을 더욱 정확하게 예측함으로써 실제 블록과 예측된 블록 간의 차이를 줄여 부호화의 효율을 높일 수 있으며 그를 통해 압축 효율을 향상시킬 수 있는 효과를 발생하는 매우 유용한 발명이다.As described above, according to an exemplary embodiment of the present invention, in generating a subsample prediction value of an image, encoding is applied to an interpolation that minimizes a difference between an actual component and a predicted component and encoding / decoding of the image using the encoding. By predicting the current block more accurately, it is possible to reduce the difference between the actual block and the predicted block, thereby increasing the efficiency of encoding and thereby generating an effect of improving the compression efficiency.
CROSS-REFERENCE TO RELATED APPLICATIONCROSS-REFERENCE TO RELATED APPLICATION
본 특허출원은 2010년 10월 06일 한국에 출원한 특허출원번호 제 10-2010-0097547 호 및 2011년 07월 21일 한국에 출원한 특허출원번호 제 10-2011-0072196 호에 대해 미국 특허법 119(a)조(35 U.S.C § 119(a))에 따라 우선권을 주장하면, 그 모든 내용은 참고문헌으로 본 특허출원에 병합된다. 아울러, 본 특허출원은 미국 이외에 국가에 대해서도 위와 동일한 동일한 이유로 우선권을 주장하면 그 모든 내용은 참고문헌으로 본 특허출원에 병합된다.This patent application is filed with the patent application No. 10-2010-0097547 filed in Korea on October 06, 2010 and the patent application number 10-2011-0072196 filed in Korea on July 21, 2011. If priority is claimed under section (a) (35 USC § 119 (a)), all of that is incorporated by reference into this patent application. In addition, if this patent application claims priority for the same reason as above for a country other than the United States, all the contents thereof are incorporated into this patent application by reference.

Claims (40)

  1. 영상을 부호화/복호화하는 장치에 있어서,An apparatus for encoding / decoding a video,
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및/또는 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하고 상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화기; 및A prediction block is generated from an interpolated value of a subsample component of a reference block referred to by a motion vector of the current block using an FIR filter and / or linear interpolation, and the residual block is subtracted from the current block. An image encoder for generating a quantized frequency transform block by transforming and quantizing the residual block and encoding the bit block into a bitstream; And
    비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록을 복원하는 영상 복호화기A FIR filter generates a quantized frequency transform block from a bitstream, and inversely quantizes and inverse transforms the quantized frequency transform block to obtain a value of a subsample component of a reference block referenced by a motion vector of a current block to restore and reconstruct a residual block. And an image decoder for generating a prediction block from the interpolated value using linear interpolation, and adding the reconstructed residual block and the generated prediction block to reconstruct the current block to be reconstructed.
    를 포함하는 것을 특징으로 하는 영상 부호화/복호화 장치.Image encoding / decoding apparatus comprising a.
  2. 영상을 부호화하는 장치에 있어서,In the apparatus for encoding a video,
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하는 예측부;A prediction unit generating a prediction block from values obtained by interpolating values of the subsample components of the reference block referred to by the motion vector of the current block by using an FIR filter and linear interpolation;
    상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산부;A subtraction unit for generating a residual block by subtracting the prediction block from the current block;
    상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환부;A transformer for converting the residual block to generate a frequency transform block;
    상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화부; 및A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And
    상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화부An encoder that encodes the quantized frequency transform block into a bitstream
    를 포함하는 것을 특징으로 하는 영상 부호화 장치.An image encoding apparatus comprising a.
  3. 제 2항에 있어서,The method of claim 2,
    현재블록의 색차 성분의 움직임 벡터는,The motion vector of the color difference component of the current block is
    상기 현재 블록의 휘도 성분의 움직임 벡터 값을 이용하는 것을 특징으로 하는 영상 부호화 장치.And a motion vector value of the luminance component of the current block.
  4. 제 2항에 있어서,The method of claim 2,
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:0의 형태인 경우, 상기 참조블록의 1/2 서브샘플의 값은 FIR필터를 사용하여 생성하고, 1/4 서브샘플의 값은 FIR필터를 사용하여 선형보간하는 것을 특징으로 하는 영상 부호화 장치.When the image is in the form of 4: 2: 0, a value of 1/2 subsample of the reference block is generated using an FIR filter, and a value of 1/4 subsample is linearly interpolated using an FIR filter. And an image encoding device.
  5. 제 2항에 있어서,The method of claim 2,
    상기 예측부는 상기 영상이 4:2:0의 형태인 경우 1/8 서브샘플의 값은,The predictor may have a value of 1/8 subsample when the image is 4: 2: 0.
    정수화소와 1/4 서브샘플의 선형보간을 이용하여 생성하거나, 정수화소와 1/2 서브샘플의 선형보간을 이용하여 생성하는 것을 특징으로 하는 영상 부호화 장치.An image encoding apparatus is generated by using linear interpolation of an integer pixel and a quarter subsample or by using linear interpolation of an integer pixel and a half subsample.
  6. 제 4항 또는 5항에 있어서,The method according to claim 4 or 5,
    상기 1/2 서브샘플 성분에 대한 FIR필터는 6-탭 이상을 갖는 FIR필터인 것을 특징으로 하는 영상 부호화 장치.And an FIR filter for the 1/2 subsample component is an FIR filter having 6-taps or more.
  7. 제 4항 또는 5항에 있어서,The method according to claim 4 or 5,
    상기 1/4 서브샘플의 값은,The value of the 1/4 subsample is,
    정수화소의 값에 32를 곱한 값과 상기 1/2 서브샘플의 값을 구할때 얻은 FIR필터링 결과값을 선형 보간하여 생성하는 것을 특징으로 하는 영상 부호화 장치.And an FIR filtering result obtained when multiplying a value of an integer pixel by 32 and the value of the 1/2 subsample is generated by linear interpolation.
  8. 제 4항 또는 5항에 있어서,The method according to claim 4 or 5,
    상기 예측부는,The prediction unit,
    상기 참조블록의 1/16 서브샘플의 위치까지 보간하되, 상기 1/16 서브샘플의 위치는 선형보간을 하여 생성하는 것을 특징으로 하는 영상 부호화 장치.And interpolating up to a position of a 1/16 subsample of the reference block, wherein the position of the 1/16 subsample is generated by linear interpolation.
  9. 제 2항에 있어서,The method of claim 2,
    상기 예측부는,The prediction unit,
    상기 영상이 4:4:4의 형태인 경우, 색차성분의 1/2 서브샘플의 위치 및 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 부호화 장치.And 4. If the image is in the form of 4: 4: 4, the image encoding apparatus interpolates up to a position of a 1/2 subsample and a position of a 1/4 subsample of the chrominance component.
  10. 제 2항에 있어서,The method of claim 2,
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:2의 형태인 경우, 수평방향의 색차성분의 1/8 서브샘플의 위치까지 보간하고, 수직방향의 색차성분은 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 부호화 장치.When the image is in the form of 4: 2: 2, the image is interpolated up to the position of 1/8 subsample of the color difference component in the horizontal direction, and the color difference component in the vertical direction is interpolated up to the position of 1/4 subsample. Video encoding device.
  11. 영상을 부호화하는 장치에 있어서,In the apparatus for encoding a video,
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측부;For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction unit generating a prediction block by obtaining values of all subsamples of the subsample;
    상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산부;A subtraction unit for generating a residual block by subtracting the prediction block from the current block;
    상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환부;A transformer for converting the residual block to generate a frequency transform block;
    상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화부; 및A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And
    상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화부An encoder that encodes the quantized frequency transform block into a bitstream
    를 포함하는 것을 특징으로 하는 영상 부호화 장치.An image encoding apparatus comprising a.
  12. 영상을 복호화하는 장치에 있어서,In the apparatus for decoding an image,
    비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화부;A decoder which extracts a quantized frequency transform block from the bitstream;
    상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화부;An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block;
    상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환부;An inverse transform unit which inversely transforms the frequency transform block to restore a residual block;
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하는 예측부; 및A prediction unit generating a prediction block from values obtained by interpolating values of the subsample components of the reference block referred to by the motion vector of the current block by using an FIR filter and linear interpolation; And
    상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산부An adder configured to reconstruct the current block by adding the reconstructed residual block and the prediction block;
    를 포함하는 것을 특징으로 하는 영상 복호화 장치.Video decoding apparatus comprising a.
  13. 제 12항에 있어서,The method of claim 12,
    현재 블록의 색차 성분의 움직임 벡터는,The motion vector of the color difference component of the current block is
    상기 현재 블록의 휘도 성분의 움직임 벡터 값을 이용하는 것을 특징으로 하는 영상 복호화 장치.And a motion vector value of the luminance component of the current block.
  14. 제 12항에 있어서,The method of claim 12,
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:0의 형태인 경우, 상기 참조블록의 1/2 서브샘플의 값은 FIR필터를 사용하여 생성하고, 1/4 서브샘플의 값은 FIR필터를 사용하여 선형보간하는 것을 특징으로 하는 영상 복호화 장치.When the image is in the form of 4: 2: 0, a value of 1/2 subsample of the reference block is generated using an FIR filter, and a value of 1/4 subsample is linearly interpolated using an FIR filter. An image decoding device.
  15. 제 12항에 있어서,The method of claim 12,
    상기 예측부는 상기 영상이 4:2:0의 형태인 경우 1/8 서브샘플의 값은,The predictor may have a value of 1/8 subsample when the image is 4: 2: 0.
    정수화소와 1/4 서브샘플의 선형보간을 이용하여 생성하거나, 정수화소와 1/2 서브샘플의 선형보간을 이용하여 생성하는 것을 특징으로 하는 영상 복호화 장치.An image decoding apparatus using linear interpolation of integer pixels and quarter subsamples, or linear interpolation of integer pixels and half subsamples.
  16. 제 14항 또는 제 15항에 있어서,The method according to claim 14 or 15,
    상기 1/2 서브샘플 성분에 대한 FIR필터는 6-탭 이상을 갖는 FIR필터인 것을 특징으로 하는 영상 복호화 장치.And an FIR filter for the 1/2 subsample component is an FIR filter having 6-taps or more.
  17. 제 14항 또는 제 15항에 있어서,The method according to claim 14 or 15,
    상기 1/4 서브샘플의 값은,The value of the 1/4 subsample is,
    정수화소의 값에 32를 곱한 값과 상기 1/2 서브샘플의 값을 구할때 얻은 FIR필터링 결과값을 선형 보간하여 생성하는 것을 특징으로 하는 영상 복호화 장치.And an FIR filtering result obtained when multiplying a value of an integer pixel by 32 and the value of the 1/2 subsample is generated by linear interpolation.
  18. 제 14항 또는 제 15항에 있어서,The method according to claim 14 or 15,
    상기 예측부는,The prediction unit,
    상기 참조블록의 1/16 서브샘플의 위치까지 보간하되, 상기 1/16 서브샘플의 위치는 선형보간을 하여 생성하는 것을 특징으로 하는 영상 복호화 장치.And interpolating up to a position of a 1/16 subsample of the reference block, wherein the position of the 1/16 subsample is generated by linear interpolation.
  19. 제 12항에 있어서,The method of claim 12,
    상기 예측부는,The prediction unit,
    상기 영상이 4:4:4의 형태인 경우, 상기 색차성분의 1/2 서브샘플의 위치 및 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 복호화 장치.And when the image is in the form of 4: 4: 4, interpolating up to a position of 1/2 subsample and a position of 1/4 subsample of the chrominance component.
  20. 제 12항에 있어서,The method of claim 12,
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:2의 형태인 경우, 수평방향의 색차성분의 1/8 서브샘플의 위치까지 보간하고, 수직방향의 색차성분은 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 복호화 장치.When the image is in the form of 4: 2: 2, the image is interpolated up to the position of 1/8 subsample of the color difference component in the horizontal direction, and the color difference component in the vertical direction is interpolated up to the position of 1/4 subsample. Video decoding device.
  21. 영상을 복호화하는 장치에 있어서,In the apparatus for decoding an image,
    비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화부;A decoder which extracts a quantized frequency transform block from the bitstream;
    상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화부;An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block;
    상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환부;An inverse transform unit which inversely transforms the frequency transform block to restore a residual block;
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측부; 및For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction unit generating a prediction block by obtaining values of all subsamples of the subsample; And
    상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산부An adder configured to reconstruct the current block by adding the reconstructed residual block and the prediction block;
    를 포함하는 것을 특징으로 하는 영상 복호화 장치.Video decoding apparatus comprising a.
  22. 영상을 부호화/복호화하는 방법에 있어서,In the method of encoding / decoding an image,
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하고 상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하고 상기 잔차 블록을 변환 및 양자화하여 양자화된 주파수변환블록을 생성하고 비트스트림으로 부호화하는 영상 부호화단계; 및Generate a residual block by subtracting the value of the subsample component of the reference block referred to by the motion vector of the current block using an FIR filter and linear interpolation, and subtracting the prediction block from the current block to generate a residual block An image encoding step of transforming and quantizing the residual block to generate a quantized frequency transform block and encoding the bitstream into a bitstream; And
    비트스트림으로부터 양자화된 주파수변환 블록을 생성하고 상기 양자화된 주파수변환블록을 역 양자화 및 역변환하여 잔차 블록을 복원하고 복원할 현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하고 상기 복원되는 잔차 블록과 상기 생성된 예측 블록을 가산하여 상기 복원할 현재 블록을 복원하는 영상 복호화단계A FIR filter generates a quantized frequency transform block from a bitstream, and inversely quantizes and inverse transforms the quantized frequency transform block to obtain a value of a subsample component of a reference block referenced by a motion vector of a current block to restore and reconstruct a residual block. And generating a prediction block from the interpolated value using linear interpolation, and adding the reconstructed residual block and the generated prediction block to reconstruct the current block to be reconstructed.
    를 포함하는 것을 특징으로 하는 영상 부호화/복호화 방법.Image encoding / decoding method comprising a.
  23. 영상을 부호화하는 방법에 있어서,In the method of encoding an image,
    현재 블록을 움직임 보상하여 움직임 벡터를 구하고 상기 움직임벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하는 예측단계;A prediction step of obtaining a motion vector by motion compensating a current block and generating a prediction block from values obtained by interpolating values of subsample components of a reference block referred to by the motion vector using an FIR filter and linear interpolation;
    상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산단계;Subtracting the prediction block from the current block to generate a residual block;
    상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환단계;Transforming the residual block to generate a frequency transform block;
    상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화단계; 및A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And
    상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화단계An encoding step of encoding the quantized frequency transform block into a bitstream
    를 포함하는 것을 특징으로 하는 영상 부호화 방법.Image encoding method comprising a.
  24. 제 23항에 있어서,The method of claim 23, wherein
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:0의 형태인 경우, 상기 참조블록의 1/2 서브샘플의 값은 FIR필터를 사용하여 생성하고, 1/4 서브샘플의 값은 FIR필터를 사용하여 선형보간하는 것을 특징으로 하는 영상 부호화 방법.When the image is in the form of 4: 2: 0, a value of 1/2 subsample of the reference block is generated using an FIR filter, and a value of 1/4 subsample is linearly interpolated using an FIR filter. An image encoding method.
  25. 제 23항에 있어서,The method of claim 23, wherein
    상기 예측부는 상기 영상이 4:2:0의 형태인 경우 1/8 서브샘플의 값은,The predictor may have a value of 1/8 subsample when the image is 4: 2: 0.
    정수화소와 1/4 서브샘플의 선형보간을 이용하여 생성하거나, 정수화소와 1/2 서브샘플의 선형보간을 이용하여 생성하는 것을 특징으로 하는 영상 부호화 방법.And a linear interpolation of an integer pixel and a quarter subsample or a linear interpolation of an integer pixel and a half subsample.
  26. 제 24항 또는 제25항에 있어서,The method of claim 24 or 25,
    상기 1/2 서브샘플 성분에 대한 FIR필터는 6-탭 이상을 갖는 FIR필터인 것을 특징으로 하는 영상 부호화 방법.And a FIR filter for the 1/2 subsample component is an FIR filter having 6-taps or more.
  27. 제 24항 또는 제25항에 있어서,The method of claim 24 or 25,
    상기 1/4 서브샘플의 값은,The value of the 1/4 subsample is,
    정수화소의 값에 32를 곱한 값과 상기 1/2 서브샘플의 값을 구할때 얻은 FIR필터링 결과값을 선형 보간하여 생성하는 것을 특징으로 하는 영상 부호화 방법.And an FIR filtering result obtained by multiplying a value of an integer pixel by 32 and the value of the 1/2 subsample is generated by linear interpolation.
  28. 제 24항 또는 제25항에 있어서,The method of claim 24 or 25,
    상기 예측단계는,The prediction step,
    상기 참조블록의 1/16 서브샘플의 위치까지 보간하되, 상기 1/16 서브샘플의 위치는 선형보간을 하여 생성하는 것을 특징으로 하는 영상 부호화 방법.And interpolating up to a position of a 1/16 subsample of the reference block, wherein the position of the 1/16 subsample is generated by linear interpolation.
  29. 제 23항에 있어서,The method of claim 23, wherein
    상기 예측단계는,The prediction step,
    상기 영상이 4:4:4의 형태인 경우, 상기 색차성분의 1/2 서브샘플의 위치 및 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 부호화 방법.And when the image is in the form of 4: 4: 4, interpolating up to a position of a 1/2 subsample and a position of a 1/4 subsample of the chrominance component.
  30. 제 23항에 있어서,The method of claim 23, wherein
    상기 예측단계는,The prediction step,
    상기 영상이 4:2:2의 형태인 경우, 수평방향의 색차성분의 1/8 서브샘플의 위치까지 보간하고, 수직방향의 색차성분은 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 부호화 방법.When the image is in the form of 4: 2: 2, the image is interpolated up to the position of 1/8 subsample of the color difference component in the horizontal direction, and the color difference component in the vertical direction is interpolated up to the position of 1/4 subsample. Image coding method.
  31. 영상을 부호화하는 방법에 있어서,In the method of encoding an image,
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측단계;For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction step of generating a prediction block by obtaining values of all subsamples of the subsample;
    상기 현재 블록에서 상기 예측 블록을 감산하여 잔차 블록을 생성하는 감산단계;Subtracting the prediction block from the current block to generate a residual block;
    상기 잔차 블록을 변환하여 주파수변환블록을 생성하는 변환단계;Transforming the residual block to generate a frequency transform block;
    상기 주파수변환블록을 양자화하여 양자화된 주파수변환블록을 생성하는 양자화단계; 및A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And
    상기 양자화된 주파수변환블록을 비트스트림으로 부호화하는 부호화단계An encoding step of encoding the quantized frequency transform block into a bitstream
    를 포함하는 것을 특징으로 하는 영상 부호화 방법.Image encoding method comprising a.
  32. 영상을 복호화하는 방법에 있어서,In the method of decoding an image,
    비트스트림으로부터 양자화된 주파수변환 블록을 생성하는 복호화단계;A decoding step of generating a quantized frequency transform block from the bitstream;
    상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화단계;An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block;
    상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환단계;An inverse transform step of restoring a residual block by inversely transforming the frequency transform block;
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 서브샘플 성분의 값을 FIR필터 및 선형보간을 이용하여 보간한 값으로부터 예측블록을 생성하는 예측단계; 및A prediction step of generating a prediction block from values obtained by interpolating values of subsample components of the reference block referred to by the motion vector of the current block by using an FIR filter and linear interpolation; And
    상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산단계An addition step of reconstructing the current block by adding the reconstructed residual block and the prediction block;
    를 포함하는 것을 특징으로 하는 영상 복호화 방법.Image decoding method comprising a.
  33. 제 32항에 있어서,The method of claim 32,
    상기 예측부는,The prediction unit,
    상기 영상이 4:2:0의 형태인 경우, 상기 참조블록의 1/2 서브샘플의 값은 FIR필터를 사용하여 생성하고, 1/4 서브샘플의 값은 FIR필터를 사용하여 선형보간하는 것을 특징으로 하는 영상 복호화 방법.When the image is in the form of 4: 2: 0, a value of 1/2 subsample of the reference block is generated using an FIR filter, and a value of 1/4 subsample is linearly interpolated using an FIR filter. An image decoding method characterized by.
  34. 제 32항에 있어서,The method of claim 32,
    상기 예측부는 상기 영상이 4:2:0의 형태인 경우 1/8 서브샘플의 값은,The predictor is a value of 1/8 subsample when the image is 4: 2: 0,
    정수화소와 1/4 서브샘플의 선형보간을 이용하여 생성하거나, 정수화소와 1/2 서브샘플의 선형보간을 이용하여 생성하는 것을 특징으로 하는 영상 복호화 방법.A method of generating an image using linear interpolation of an integer pixel and a quarter subsample or a linear interpolation of an integer pixel and a half subsample.
  35. 제 33항 또는 제 34항에 있어서,The method of claim 33 or 34,
    상기 1/2 서브샘플 성분에 대한 FIR필터는 6-탭 이상을 갖는 FIR필터인 것을 특징으로 하는 영상 복호화 방법.And an FIR filter for the 1/2 subsample component is an FIR filter having 6-taps or more.
  36. 제 33항 또는 제 34항에 있어서,The method of claim 33 or 34,
    상기 1/4 서브샘플의 값은,The value of the 1/4 subsample is,
    정수화소의 값에 32를 곱한 값과 상기 1/2 서브샘플의 값을 구할때 얻은 FIR필터링 결과값을 선형 보간하여 생성하는 것을 특징으로 하는 영상 복호화 방법.And generating a linear interpolation of the FIR filtering result obtained when multiplying the integer pixel value by 32 and the 1/2 subsample value.
  37. 제 33항 또는 제 34항에 있어서,The method of claim 33 or 34,
    상기 예측단계는,The prediction step,
    상기 참조블록의 1/16 서브샘플의 위치까지 보간하되, 상기 1/16 서브샘플의 위치는 선형보간을 하여 생성하는 것을 특징으로 하는 영상 복호화 방법.And interpolating up to a position of a 1/16 subsample of the reference block, wherein the position of the 1/16 subsample is generated by linear interpolation.
  38. 제 32항에 있어서,The method of claim 32,
    상기 예측단계는,The prediction step,
    상기 영상이 4:4:4의 형태인 경우, 상기 색차성분의 1/2 서브샘플의 위치 및 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 복호화 방법.And when the image is in the form of 4: 4: 4, interpolating up to a position of 1/2 subsample and a position of 1/4 subsample of the chrominance component.
  39. 제 32항에 있어서,The method of claim 32,
    상기 예측단계는,The prediction step,
    상기 영상이 4:2:2의 형태인 경우, 수평방향의 색차성분의 1/8 서브샘플의 위치까지 보간하고, 수직방향의 색차성분은 1/4 서브샘플의 위치까지 보간하는 것을 특징으로 하는 영상 복호화 방법.When the image is in the form of 4: 2: 2, the image is interpolated up to the position of 1/8 subsample of the color difference component in the horizontal direction, and the color difference component in the vertical direction is interpolated up to the position of 1/4 subsample. Image Decoding Method.
  40. 영상을 복호화하는 방법에 있어서,In the method of decoding an image,
    비트스트림으로부터 양자화된 주파수변환 블록을 추출하는 복호화단계;Decoding the quantized frequency transform block from the bitstream;
    상기 양자화된 주파수변환블록을 역 양자화하여 주파수변환블록을 복원하는 역양자화단계;An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block;
    상기 주파수변환블록을 역 변환하여 잔차 블록을 복원하는 역변환단계;An inverse transform step of restoring a residual block by inversely transforming the frequency transform block;
    현재 블록의 움직임 벡터에 의해 참조되는 참조블록의 1/2 서브샘플에 대하여 주변의 정수화소값에 필터링 계수를 적용한 1/2샘플 고정밀도를 구하고 상기 1/2샘플 고정밀도를 이용하여 상기 참조블록의 모든 서브샘플의 값을 구하여 예측블록을 생성하는 예측단계; 및For a 1/2 subsample of a reference block referred to by the motion vector of the current block, a 1/2 sample precision obtained by applying a filtering coefficient to an integer pixel value of a neighbor is obtained and the 1/2 sample high precision is used for the reference block. A prediction step of generating a prediction block by obtaining values of all subsamples of the subsample; And
    상기 복원되는 잔차 블록과 상기 예측 블록을 가산하여 상기 현재 블록을 복원하는 가산단계An addition step of reconstructing the current block by adding the reconstructed residual block and the prediction block;
    를 포함하는 것을 특징으로 하는 영상 복호화 방법.Image decoding method comprising a.
PCT/KR2011/007418 2010-10-06 2011-10-06 Method and apparatus for encoding/decoding video using high-precision filter WO2012047047A2 (en)

Priority Applications (5)

Application Number Priority Date Filing Date Title
CN201180054874.XA CN103210649B (en) 2010-10-06 2011-10-06 The method and apparatus using high-precision filter coding/decoding video
US13/857,708 US9420281B2 (en) 2010-10-06 2013-04-05 Method and apparatus for encoding/decoding video using high-precision filter
US15/031,366 US9706222B2 (en) 2010-10-06 2016-04-22 Method and apparatus for encoding/decoding video using high-precision filter
US15/031,378 US9602834B2 (en) 2010-10-06 2016-04-22 Method and apparatus for encoding/decoding video using high-precision filter
US15/371,215 US10158880B2 (en) 2010-10-06 2016-12-07 Method and apparatus for encoding/decoding video using high-precision filter

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
KR20100097547 2010-10-06
KR10-2010-0097547 2010-10-06
KR10-2011-0072196 2011-07-21
KR1020110072196A KR101341993B1 (en) 2010-10-06 2011-07-21 Method and Apparatus for Image Encoding/Decoding Using High Resolution Filter

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/857,708 Continuation US9420281B2 (en) 2010-10-06 2013-04-05 Method and apparatus for encoding/decoding video using high-precision filter

Publications (2)

Publication Number Publication Date
WO2012047047A2 true WO2012047047A2 (en) 2012-04-12
WO2012047047A3 WO2012047047A3 (en) 2012-06-28

Family

ID=45928243

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2011/007418 WO2012047047A2 (en) 2010-10-06 2011-10-06 Method and apparatus for encoding/decoding video using high-precision filter

Country Status (1)

Country Link
WO (1) WO2012047047A2 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107211150A (en) * 2015-01-16 2017-09-26 微软技术许可有限责任公司 Dynamic updates quality to higher chroma samples rate
US10523953B2 (en) 2012-10-01 2019-12-31 Microsoft Technology Licensing, Llc Frame packing and unpacking higher-resolution chroma sampling formats
CN113302934A (en) * 2019-12-23 2021-08-24 Oppo广东移动通信有限公司 Image prediction method, encoder, decoder, and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004015999A1 (en) * 2002-08-08 2004-02-19 Matsushita Electric Industrial Co., Ltd. Moving picture encoding method and decoding method
KR20100030119A (en) * 2008-09-09 2010-03-18 에스케이 텔레콤주식회사 Inter prediction apparatus of considering rotational motion and video encoding/decoding method and apparatus using same
WO2010044569A2 (en) * 2008-10-16 2010-04-22 에스케이텔레콤 주식회사 Method and apparatus for generating a reference frame, and method and apparatus for image encoding/decoding using same

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004015999A1 (en) * 2002-08-08 2004-02-19 Matsushita Electric Industrial Co., Ltd. Moving picture encoding method and decoding method
KR20100030119A (en) * 2008-09-09 2010-03-18 에스케이 텔레콤주식회사 Inter prediction apparatus of considering rotational motion and video encoding/decoding method and apparatus using same
WO2010044569A2 (en) * 2008-10-16 2010-04-22 에스케이텔레콤 주식회사 Method and apparatus for generating a reference frame, and method and apparatus for image encoding/decoding using same

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10523953B2 (en) 2012-10-01 2019-12-31 Microsoft Technology Licensing, Llc Frame packing and unpacking higher-resolution chroma sampling formats
CN107211150A (en) * 2015-01-16 2017-09-26 微软技术许可有限责任公司 Dynamic updates quality to higher chroma samples rate
CN107211150B (en) * 2015-01-16 2020-02-28 微软技术许可有限责任公司 Method, computer system and computer readable medium for dynamically updating quality to higher chroma sampling rate
CN113302934A (en) * 2019-12-23 2021-08-24 Oppo广东移动通信有限公司 Image prediction method, encoder, decoder, and storage medium
CN113302934B (en) * 2019-12-23 2022-10-18 Oppo广东移动通信有限公司 Image prediction method, encoder, decoder, and storage medium

Also Published As

Publication number Publication date
WO2012047047A3 (en) 2012-06-28

Similar Documents

Publication Publication Date Title
WO2013070006A1 (en) Method and apparatus for encoding and decoding video using skip mode
WO2013002549A2 (en) Method and apparatus for coding/decoding image
WO2012077960A2 (en) Method and device for encoding/decoding image by inter prediction using random block
WO2013005941A2 (en) Apparatus and method for coding and decoding an image
WO2009157669A2 (en) Intra prediction method and apparatus, and image encoding/decoding method and apparatus using same
KR101648058B1 (en) Method and Apparatus for Image Encoding/Decoding Using High Resolution Filter
WO2011145819A2 (en) Image encoding/decoding device and method
WO2013002550A2 (en) Method and apparatus for coding/decoding through high-speed coding unit mode decision
WO2010027182A2 (en) Method and device for image encoding/decoding using arbitrary pixels in a sub-block
WO2012096550A2 (en) Method and device for encoding/decoding image using bi-directional intra prediction
WO2009157665A2 (en) Intra prediction method and apparatus using block transform, and image encoding/decoding method and apparatus using the same
WO2010044569A2 (en) Method and apparatus for generating a reference frame, and method and apparatus for image encoding/decoding using same
WO2012093854A2 (en) Method and device for encoding and decoding by using parallel intra-prediction by a coding unit
WO2012015275A2 (en) Method and device for image encoding/decoding using block split prediction
WO2013109039A1 (en) Image encoding/decoding method and apparatus using weight prediction
WO2010047499A2 (en) Video encoding/decoding apparatus, de-blocking filter and filtering method based on intra-prediction directions for same, and recording media
WO2012091461A2 (en) Method and device for encoding/decoding image using feature vectors of surrounding blocks
WO2012033344A2 (en) Method and apparatus for encoding/decoding images using the effective selection of an intra-prediction mode group
WO2012011672A2 (en) Method and device for encoding/decoding image using extended skip mode
WO2012077959A2 (en) Method and device for encoding/decoding image in spatial domain for noise component
WO2012021040A2 (en) Method and device for encoding/decoding image having removable filtering mode
WO2012081917A2 (en) Method and apparatus for generating coded motion information/recovering motion information using motion information integration, and image encoding/decoding method and apparatus using same
WO2011052990A2 (en) Method and apparatus for encoding/decoding images based on adaptive resolution
WO2018074626A1 (en) Video coding method and apparatus utilizing adaptive interpolation filter
WO2012047047A2 (en) Method and apparatus for encoding/decoding video using high-precision filter

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11830932

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase in:

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/07/13)

122 Ep: pct application non-entry in european phase

Ref document number: 11830932

Country of ref document: EP

Kind code of ref document: A2