CN104243996A - In-loop processor, selective pixel difference value compensation method and encoding system - Google Patents
In-loop processor, selective pixel difference value compensation method and encoding system Download PDFInfo
- Publication number
- CN104243996A CN104243996A CN201310236938.9A CN201310236938A CN104243996A CN 104243996 A CN104243996 A CN 104243996A CN 201310236938 A CN201310236938 A CN 201310236938A CN 104243996 A CN104243996 A CN 104243996A
- Authority
- CN
- China
- Prior art keywords
- bit rate
- offset
- coding
- value
- distortion value
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Provided are an in-loop processor, a selective pixel difference value compensation method and an encoding system. Image data obtained before encoding and reconstructed image data obtained after encoding serve as references in the method, a plurality of compensation values are generated according to difference states of pixels in blocks of the two kinds of the image data, the most suitable compensation value is selectively chosen according to preset conditions to serve as an inconsistent signal, and therefore an error of the reconstructed image data obtained after encoding is compensated.
Description
Technical field
The present invention relates to a kind of image processing apparatus and method; In particular to one optionally pixel value difference compensation method and device.
Background technology
In the video coding in modern age (H.264, HEVC, AVS), contain filter (in-loop filter) in a loop, in order to the picture of having encoded is got rid of partial noise, be then referred to as deblocking filter (De-blocking filter) as what get rid of blocking artifact (blocking effect).And the result be filtered, the video pictures that can be stored for follow-up input is referenced in time encoding, so be called the filter in loop.
But the view data of filter often only after reference encoder is predicted, as inter-picture prediction (Inter-prediction) in loop.Due to filter reference in only loop according to be only the numerical value of prediction not with reference to the data of original image, therefore cannot accurately revise the difference of having encoded between picture and raw frames.
Summary of the invention
An object of the present invention is providing one optionally pixel value difference compensation method and device, view data before the method reference encoder and the view data after encoding, multiple offset is produced according to pixel difference state in the square of this two view data, optionally select most suitable offset as an inconsistent signal according to pre-conditioned, to compensate the error of the rear reconstructed image data of coding.
One embodiment of the invention provide one optionally pixel value difference compensation method, include the following step: first, and receive the rear reconstructed image data of a coding, after coding, reconstructed image data divides into multiple square.Then, the front view data of a coding is received.Afterwards, according to the pixel difference status of reconstructed image data same position after view data and coding before encoding of each square, statistical discrepancy state produces multiple offset, optionally chooses an offset to produce an inconsistent signal.
Another embodiment of the present invention provides processing unit in a kind of loop, is applicable to a coded system, includes a difference compensation arrangement.After difference compensation arrangement receives a coding reconstructed image data and an encode before view data, and the picture of rear reconstructed image data of encoding divides into multiple square.Difference compensation arrangement, according to pixel view data and the difference status of the rear reconstructed image data same position of coding before encoding of each square, produces an inconsistent signal.Wherein, the difference status of each square of difference compensation arrangement comprises the edge of the rear reconstructed image data of coding, multiple classification numerical value, the multiple brightness variable condition section numerical value with angle variable condition, optionally select the part classifying numerical value of multiple classification numerical value, with the partial sector numerical value of the plurality of brightness variable condition section numerical value as offset, produce this inconsistent signal.
Another embodiment of the present invention provides a kind of coded system, includes a motion compensating device and a difference compensation arrangement.Motion compensating device is in order to the displacement error compensating last the rear reconstructed image data of coding and input at present between picture (before current coding view data).Difference compensation arrangement is coupled between conversion quantization device and motion compensating device, to receive after current encoder view data before reconstructed image data and current encoder.Wherein, after current encoder, the picture of reconstructed image data divides into multiple square, difference compensation arrangement, according to the pixel of each square multiple difference status of reconstructed image data same position after view data and current encoder before current encoder, produces an inconsistent signal.Wherein, difference compensation arrangement optionally selects the numerical value of the part variation state of multiple difference status as offset, produces above-mentioned inconsistent signal.
Processing unit and coded system in the optionally pixel value difference compensation arrangement of the embodiment of the present invention and method, loop, reconstructed image data view data front with coding after reference encoder, and in units of square, various modes assessment is used to each square, judge to use which kind of pattern can improve picture quality under the condition not increasing or increase a little bit rate.And the ratio of the recruitment of bit rate can be less than the ratio of the picture quality of raising in one embodiment.
Accompanying drawing explanation
Figure 1A is the schematic diagram showing coded system according to an embodiment of the invention.
Figure 1B is the schematic diagram showing coded system according to another embodiment of the present invention.
Fig. 1 C is the schematic diagram showing according to an embodiment of the invention optionally pixel value difference compensation arrangement.
Fig. 2 A is the flow chart showing optionally pixel value difference compensation method according to an embodiment of the invention.
Fig. 2 B is the flow chart showing step S203 according to an embodiment of the invention.
Fig. 3 A is the schematic diagram showing spike and low ebb according to an embodiment of the invention.
Fig. 3 B is the schematic diagram showing angle classification according to an embodiment of the invention.
Fig. 3 C is the schematic diagram showing section distribution according to an embodiment of the invention.
[symbol description]
100 coded systems
101 difference compensation arrangements
102 coding controllers
103 conversion quantization devices
103a conversion/convergent-divergent/quantifying unit
103b convergent-divergent and investigation converting unit
Unit is preset in 104 pictures
105 motion compensations and estimation unit
105a motion compensation unit
105b moves estimation unit
106 random degree encoders
107 remove block effect filter
101a lines mode processing units
101b smooth mode processing unit
101c judging unit
Embodiment
Figure 1A shows the schematic diagram of the coded system 100 of one embodiment of the invention.Coded system 100 include one optionally pixel value difference compensation arrangement (Selective Pixel DisparityCompensation, SPDC) 101, one coding controller (Coder control) 102, change quantization device 103, intra-frame prediction unit (Intra-frame Prediction) 104, motion compensation and estimation unit 105 and a random degree encoder (Entropy Coding) 106.One embodiment, conversion quantization device 103 includes one conversion/convergent-divergent/quantifying unit (Transform/Scal./Quant.) 103a and convergent-divergent and investigation converting unit (Scaling & Inv.Transform) 103b.One motion compensation and estimation unit 105 include a motion compensation unit (Motion Compensation) 105a and and move estimation unit (Motion Estimation) 105b.
(input) view data BCD before coded system 100 received code, reconstructed image data ACD after encoding will be produced after view data BCD coding before encoding, again after the process such as motion compensation, via random degree encoder 106 output image data OS, to export a code translator decoding to and to show image.In one embodiment, decoding device such as can be the decoding device of the device (as contact surface plate, hand-held moving device, mobile glasses etc.) being arranged at a TV (TV) or various displayable image.
Should be noted, in one embodiment, in this figure, signal or data symbol CD represent a control data (Control Data), QTC represents a quantification transmission coefficient (Quant.Transf.coeffs), OV represents that output image (Output Video) represents Mobile data (Motion Data) for coding side reference, MD.
In one embodiment, difference compensation arrangement 101 couples conversion quantization device 103 and motion compensation unit 105a.Reconstructed image data ACD view data BCD front with coding after difference compensation arrangement 101 received code, view data is divided into multiple square (Blocks), for each square before encoding, the state difference of the pixel of rear view data BCD, ACD, calculate offset Com(not shown), to produce inconsistent signal (Disparity Signal) DS to random degree encoder 106, extra information is provided to improve the quality of image.
Should be noted, the difference compensation arrangement 101 of the embodiment of the present invention is different from known technology, and difference compensation arrangement 101 compensates according to the raw image data before coding in addition, and the non-data as known technology only reference prediction are as the foundation of offset.Moreover difference compensation arrangement 101 can optionally find out optimal offset Com according to the pixel classifications of each square and state characteristic, is used as inconsistent signal DS.
Should be noted, difference compensation arrangement 101 of the present invention can be arranged in various coded system, is not limited to the framework of above-described embodiment.Moreover except difference compensation arrangement 101 of the present invention, those skilled in the art, should understand operation principles and the method for the framework of above-described embodiment, repeat no more its content.
In one embodiment, as shown in Figure 1B, coded system 100 also comprises block effect filter (De-blocking Filter) 107.After the difference compensation arrangement 101 of the present embodiment can be arranged at block effect filter 107, receive and remove view data BCD before the output signal of block effect filter 107 and coding, to export inconsistent signal DS to random degree encoder 106.
In one embodiment, when coded system 100 does not arrange block effect filter 107, difference compensation arrangement 101 can be used as filter (In-loop filter) in processing unit in loop (In-loop processing device) or loop and uses, and replaces and removes block effect filter 107.
The schematic diagram of processing unit-difference compensation arrangement 101 in the loop that Fig. 1 C shows one embodiment of the invention.Reconstructed image data ACD view data BCD front with coding after difference compensation arrangement 101 received code, according to pixel view data BCD and the diversity factor of the rear reconstructed image data ACD same position pixel of coding before encoding of each square in view data, produce an inconsistent signal DS.
As shown in Figure 1 C, difference compensation arrangement 101 comprises a lines pattern (Texture Mode) processing unit 101a, a smooth mode (Flat Mode) processing unit 101b and judging unit 101c.
Lines mode processing units 101a is by each pixel basis edge (Edge) of each square of reconstructed image data BCD after coding and the state classification of height (Peak and Valley), and the number of pixels calculating each classification produces a differential counting value count_t.The pixel value summation of each classification in reconstructed image data ACD after calculation code, classify after producing a coding total value sum(Coded) _ t.The pixel value summation of each classification in view data BCD before calculation code, classify before producing coding total value sum(Org) _ t; Utilize random degree encoder 106 to calculate the pixel of each classification, obtain the bit rate (Bit rate) of this classification.The coding of each this differential counting value count_t classified, each classification is classified total value sum(Coded afterwards) _ t, each classification coding before classification total value sum(Org) bit rate of _ t and this classification substitutes into a bit rate and distortion value equation, and substitute into the offset preset, to calculate the first bit rate and distortion value RDt and to find out each offset of classifying (Compensation).Then, select the offset of the best producing minimal bit rate and distortion value (Rate-Distortion) RDt as the first offset Com1.
The distributed area of whole pixel basis brightness of each square of reconstructed image data BCD after coding is divided into multiple brightness section by smooth mode processing unit 101b, calculates the number of pixels of each section, produces a sector counter values count_f; The pixel value summation of each section in reconstructed image data ACD after calculation code, produces section total value sum(Coded after coding) _ f; The pixel value summation of each section in view data BCD before calculation code, produces coding front section total value sum(Org) _ f; Random degree encoder 106 is utilized to calculate the bit rate of each section pixel; Then, in multiple section, select continuous multiple section, by section total value sum(Coded after the sector counter values count_f of each section in the plurality of section, coding) _ f, coding front section total value sum(Org) _ f and this bit rate substitute into a bit rate and distortion value equation, and substitute into the offset preset, to calculate the second bit rate and distortion value RDf and to find out the offset (Compensation) of each section.Then, select the offset of the best producing minimal bit rate and distortion value RDf as the second offset Com2.
Judging unit 101c compares the first bit rate and distortion value RDt and the second bit rate and distortion value RDf, when the first bit rate and distortion value RDt are less than the second bit rate and distortion value RDf, and the first bit rate and distortion value RDt are when being less than zero, export the first offset Com1, using as inconsistent signal DS.And when the second bit rate and distortion value RDf are less than the first bit rate and distortion value RDt, and when the second bit rate and distortion value RDf are less than zero, export the second offset Com2, using as inconsistent signal DS.Should be noted, bit rate and distortion value the smaller the better, and must zero be less than, just can promote image quality not affecting under usefulness.When first, second bit rate and distortion value RDt, RDf are all more than or equal to zero, adopt " shut " mode" (Off Mode), do not compensate reconstructed image data ACD after coding, after directly encoding, reconstructed image data ACD exports.
As shown in Figure 1A, 1B, the random degree encoder 106 of coded system 200 receives inconsistent signal DS, according to reconstructed image data ACD after inconsistent signal compensation coding, to export reconstructed image data to decoder after the coding after compensation.
Next, describe the present invention's optionally pixel value difference compensation method in detail, please also refer to Fig. 1 C, 2A, 2B, 3A, 3B and 3C.
As shown in Fig. 2 A, 2B, Fig. 2 A, 2B show the flow chart of the present invention's optionally pixel value difference compensation method.The method comprises the following step:
Step S201: start.
Step S202: the reconstruction image frame (i.e. coding after reconstructed image data ACD) after coding is cut into N × N's such as 6464) square, difference compensation arrangement 101 receives the square of N × N.Wherein, N is positive integer.
Step S203: for each square finds this square and uncoded front image frame same position, calculate pixel reconstruction image frame in encoded and the diversity factor of uncoded front image frame of same position, optionally find out the values of disparity between pixel under the classification of various characteristic, and select most suitable values of disparity, inconsistent signal DS is produced, to compensate (disparity compensation) this diversity factor according to values of disparity.Wherein, the generation of diversity factor can use multiple (such as Three models) to go assessment to each square, this Three models includes a lines pattern (Texture Mode), a smooth mode (Flat Mode), a " shut " mode" (OFF Mode), to determine to use which kind of pattern can increase picture quality under the condition increasing reasonable bit rate.In addition, the bit that this inconsistent signal additionally sent increases, is set as causing the size making bit rate increase under the condition of same picture quality, such as (Bjontegaard Delta Bit-Rate) BDBR<0.
Please refer to Fig. 2 B, step S203 also comprises the following step:
Step S2031: a lines pattern is provided, lines schema processing unit 101a, pixel (pixel) being classified with the concept of edge (edge), such as, is following five kinds of states by pixel classifications: (1) edge and be low ebb (valley) pixel, (2) non-edge and be low ebb pixel, (3) edge and be spike (peak) pixel, (4) non-edge and be spike (peak) pixel, all non-more than (5).Then, add up the pixel of (1) ~ (4) kenel, pixel value sum, preceding pixel value of encoding sum after finding out the quantity (count) of each kenel group, coding, to find out the offset of this four kenels group the most applicable in single square.Should be noted, the spike (Peak) at edge and low ebb (Valley), as shown in Figure 3A.
In one embodiment, as shown in Figure 3 B, due to conventional angle have level, vertical, Oblique 45 Degree, oblique 135 degree, so at lines mode step S2031, understand the style (pattern) of four kinds of different angles for this reason, statistics out of the ordinary, by a bit rate of pixel value sum, preceding pixel value of encoding sum substitution lines mode processing units 101a after the quantity (Count) of each kenel group found out, coding and distortion value (RD) equation:
First, as Fig. 3 B(A) style 1 shown in, when the angle of pixel detection be horizontal spike or low ebb time, the edge of vertical direction of now arranging in pairs or groups, is divided into (1) vertical edge and horizontal spike, (2) non-perpendicular edge and horizontal spike, (3) vertical edge and horizontal low ebb, (4) non-perpendicular edge and horizontal low ebb by pixel.Then, search each pixel in this square, to find out the quantity (Count) of these four kinds classification difference, to calculate square whole pixel value sum (sum(Coded) of these four kinds classification respectively reconstructed image data ACD in encoded), calculate the whole pixel value sum (sum(Org) of square of these four kinds classification difference view data BCD before encoding).Then, by i.e. (1), (2), (3), (4) of each classification i() count, sum(Coded), sum(Org) substitute into equation RDt, utilize the offset Com(compesations preset) calculate bit rate and distortion value RDt, to find out most suitable offset.For example, in style 1, the Count value that (1) vertical edge and horizontal spike calculate is 100 pixels, its substitute into equational predesigned compensation value test after the offset obtained for sending be 3; (2) the Count value of non-perpendicular edge and horizontal spike is 5, its substitute into equational predesigned compensation value test after the offset obtained for sending be 0; (3) the Count value that vertical edge and horizontal low ebb calculate is 100, its substitute into equational predesigned compensation value test after the offset obtained for sending be 2; (4) the Count value that non-perpendicular edge and horizontal low ebb calculate is 10, its substitute into equational predesigned compensation value test after the offset obtained for sending be 1.Then, due to offset 3 substitute into equation after the RDt value that obtains minimum, therefore select offset 3.
Should be noted, in one embodiment, under classification (1) vertical edge of style 1 and the state of horizontal spike and (3) vertical edge and horizontal low ebb, in order to allow vertical edge can obtain preferably offset, can allow to select positive and negative number to mingle (Sign) or be full the offset of positive sign (Unsign); Relatively, under the state of classification (2) the non-perpendicular edge of style 1 and horizontal spike and (4) non-perpendicular edge and horizontal low ebb, in order to obtain lower bit rate, then do not allow the offset selecting sign, the offset of positive number (Unsign) can only be selected.In another embodiment, those skilled in the art, can the rest may be inferred determines that the classification setting compensation value of style 2 ~ 4 selects mode.
As Fig. 3 B(B) style 2 shown in, when the direction of pixel detection be vertical spike or low ebb time, pixel, to edge, is divided into (1) horizontal edge and vertical spike, (2) non-horizontal edge and vertical spike, (3) horizontal edge and vertical low ebb, (4) non-horizontal edge and vertical low ebb by level of now arranging in pairs or groups.Then, search each pixel in this square, to find out quantity (count), each whole pixel value sum (sum(Coded) of classifying of this square in reconstructed image data ACD after calculation code of these four kinds classification), each whole pixel value sum (sum(Org) of classifying of this square in view data BCD before calculation code).Then, an offset is calculated with above-mentioned style 1.
As Fig. 3 B(C) style 3 shown in, when the direction of pixel detection be 45 degree of spikes or low ebb time, now to arrange in pairs or groups horizontal or vertical direction edge, pixel is divided into (1) level or vertical edge and 45 degree of spikes, (2) non-horizontal or vertical edges and 45 degree of spikes, (3) level or vertical edges and 45 degree of low ebbs, (4) non-horizontal or vertical edges and 45 degree of low ebbs.Then, search each pixel in this square, to find out quantity (count), each whole pixel value sum (sum(Coded) of classifying of this square in reconstructed image data ACD after calculation code of these four kinds classification), each whole pixel value sum (sum(Org) of classifying of this square in view data BCD before calculation code).Then, an offset is calculated with above-mentioned style 1.
As Fig. 3 B(D) style 4 shown in, when the direction of pixel detection be 135 degree of spikes or low ebb time, now collocation horizontal or vertical direction edge, pixel is divided into (1) level or vertical edge and 135 degree of spikes, (2) non-horizontal or vertical edges and 135 degree of spikes, (3) level or vertical edges and 135 degree of low ebbs, (4) non-horizontal or vertical edge and 135 degree of low ebbs.Then, search each pixel in this square, to find out quantity (count), each whole pixel value sum (sum(Coded) of classifying of this square in reconstructed image data ACD after calculation code of these four kinds classification), each whole pixel value sum (sum(Org) of classifying of this square in view data BCD before calculation code).Then, an offset is calculated with above-mentioned style 1.
Mode according to this, lines mode processing units 101a can find out the best angle obtaining most suitable offset under four kinds of angles (level, vertical, 45 degree, 135 degree) in the direction that element detects, four offset Com(Compensations that the style optionally sending best angle is numbered and had).Such as, suppose that best angle is style (2) and four offsets are respectively 0,1,3,2.
Step S2032 a: smooth mode (Flat Mode) is provided, in this step, smooth mode processing unit 101b is with the concept of rectangular statistical chart (histogram), the pixel of each square is divided into multiple brightness section, for each section finds out the pixel value sum of each section in the pixel value sum of each section in the quantity (count) of group, the rear reconstructed image data ACD that encodes, the front view data BCD that encodes.For the section having statistic in single square, find out optimal offset in single square.
In one embodiment, as shown in Figure 3 C, brightness is divided into 32 rank (divided by 8, divisor 8, for arrange numerical value arbitrarily, the present invention is not limited thereto the square divisor therewith of size in the maximum brightness 256 of pixel) by smooth mode processing unit 101b.Then, a square is added up, by brightness binary numeral Bin 0 progressively add up to Bin32.Note that binary numeral Bin3 in this figure, 6,7,8,12,14,15 sides have cumulative statistics numerical value, for these have the Bin of cumulative statistics numerical value to redefine address, are newly defined as Bin0,3,4,5,9,11,12 respectively.Smooth mode processing unit 101b is in the bin having value, select continuous 4, as in legend 0,3,4,5 or 3,4,5,9 or 4,5,9,11 etc., find out one group at following bit rate and distortion value equation RDf...(2) in, the combination of optimum value can be obtained.Suppose original Bin6,7,8, the RDf value of 12 is 500, than original Bin3,6,7, the RDf value 1000 of 8 is low, then smooth mode processing unit 101b can select continuous 4 Bin6,7,8,12 having a value by Bin6, re-addresses the order of Bin6 into Bin3.Then, by the binary numeral index (Bin index) of the initial Bin of the combination of optimum value and four Com2(Compensations) coding, in one embodiment such as (3;-1 ,-2,1,2), send, wherein 3 is index, and-1 ,-2,1,2 is four Com2 codings.Should be noted, send index Bin3 and represent that three indexes continued are sequentially Bin4, Bin5, Bin9, and Bin4, Bin5, Bin9 also have each self-corresponding four Com2 values, these four Com2 values can be selected from-1 ,-2,1,2 in one embodiment.
Step S2033: judge whether RDt is less than RDf, if so, enters step S2034; If not, step S2035 is entered.
Step S2034: judge whether RDt is less than 0, if so, enters step S2036; If not, step S2037 is entered.Wherein, when the value of RDt is less than 0, represent that adding this offset can benefit to picture quality lifting; And when the value of RDt is greater than 0, represent that adding this offset is no advantage to picture quality lifting.
Step S2035: judge whether RDf is less than 0, if so, enters step S2038; If not, step S2037 is entered.Wherein, when the value of RDf is less than 0, represent that adding this offset can benefit to picture quality lifting; And when the value of RDf is greater than 0, represent that adding this offset is no advantage to picture quality lifting.
Step S2036: the offset Com1 that employing lines mode computation goes out is to produce inconsistent signal DS.
Step S2037: adopt " shut " mode", any offset of not bothering to see me out, maintains the state after this block encription.
Step S2038: the offset Com2 adopting level and smooth lines mode computation to go out is to produce inconsistent signal DS.
Step S204: terminate.
Please note, sampling self adaptation skew (the Sample Adaptive Offset of known technology video coding HEVC, SAO) in, be divided into " shut " mode" (Off Mode), edge difference (Edge Offset, EO), frequency range difference (Band Offset, BO) Three models, wherein, EO pattern comprises low ebb (Valley), recessed (concave), convex (convex), spike (Peak) four kinds of patterns, level and smooth and marginal zone are all included by this statistical, edge (Edge) does not process especially, moreover the also number of having selected (Sign) under which state of indefinite, which selects the offset without number (Unsign), and the statistical of BO, then unspecial in there being the section of statistic to process, all can be used regardless of the section with or without statistic.And the embodiment of the present invention, then optionally can find out most suitable statistics section, to produce most suitable offset.
When encoding, the inconsistent signal DS that optionally pixel value difference compensation arrangement 101 of the present invention produces includes mode index (Mode index), style index (Pattern index) and offset Com(Compensations).In one embodiment, mode index can be: " shut " mode" (0), lines pattern (1), smooth mode (2).Due in video pictures, have relevance between adjacent picture to exist, smooth area or lines district all can a slices continuously, so when encoding, random degree coding (context adaptive entropy coding) mode of context adaptive can be adopted, utilize the mode index of last square to predict the mode index of current square, increase coding usefulness.And the offset of style index in lines pattern and the binary numeral index (Bin index) of smooth mode, each pattern, the mode of lossless coding (lossless) can be used to encode.Should be noted, it will be understood by a person skilled in the art that above-mentioned lossless signal coding and lossless coding mode, repeat no more.
The optionally pixel value difference compensation arrangement of the embodiment of the present invention and method, view data (original image) and in units of square before reconstructed image data and coding after reference encoder, various modes assessment is used to each square, judges to use which kind of pattern can increase picture quality under the condition not increasing bit rate.
Claims (19)
1. a processing unit in loop, is applicable to a coded system, and wherein, in described loop, processing unit comprises:
One difference compensation arrangement, after described difference compensation arrangement receives a coding, reconstructed image data and is encoded front view data, after described coding, the picture of reconstructed image data divides into multiple square, described difference compensation arrangement, according to the pixel of each described square difference status of the same position of reconstructed image data after view data and described coding before described coding, produces an inconsistent signal;
Wherein, the described difference status of each described square of described difference compensation arrangement comprises edge and multiple classification numerical value of angle variable condition, multiple brightness variable condition section numerical value of reconstructed image data after described coding, described difference compensation arrangement optionally selects the partial sector numerical value of the part classifying numerical value of described multiple classification numerical value and described multiple brightness variable condition section numerical value as offset, produces described inconsistent signal.
2. processing unit in loop according to claim 1, wherein, the motion compensation unit and that described difference compensation arrangement is coupled to described coded system quantizes between converting unit.
3. processing unit in loop according to claim 1, wherein, one random degree encoder of described coded system according to reconstructed image data after coding described in described inconsistent signal compensation, to export reconstructed image data after the described coding after compensating to a decoder for decoding.
4. processing unit in loop according to claim 1, wherein, described difference compensation arrangement comprises:
One lines mode processing units, described lines mode processing units, by each pixel basis edge of described square and height state classification, calculates the number of pixels of each classification and produces a differential counting value; Calculate the pixel value summation of each described classification in reconstructed image data after described coding, total value of classifying after producing a coding; Calculate the pixel value summation of each described classification in view data before described coding, produce classification total value before coding; Each described classified pixels is calculated the bit rate of described classification by a random degree encoder; The described offset of each described classification and a bit rate and distortion value is produced according to classify afterwards to classify before total value, described coding total value and described bit rate of the described differential counting value of each described classification, described coding, size according to described bit rate and distortion value selects offset described in as one first offset, and described bit rate corresponding to described first offset and distortion value are one first bit rate and distortion value.
5. processing unit in loop according to claim 4, wherein, described difference compensation arrangement comprises:
Whole pixel regions of described square are divided into multiple brightness section, calculate the number of pixels of section described in each, produce a sector counter values by one smooth mode processing unit; Calculate the pixel value summation of each described section in reconstructed image data after described coding, produce section total value after a coding; Calculate the pixel value summation of each described section in view data before described coding, produce a coding front section total value; The pixel of each described section is utilized to calculate the bit rate of described section by described random degree encoder; Select the multiple section of continuous print, the described offset of each described section and a bit rate and distortion value is produced according to section total value, described coding front section total value and described bit rate after the described sector counter values of described section each in described multiple section, described coding, size according to described bit rate and distortion value selects offset described in as one second offset, and the described bit rate of described second offset and distortion value are the second bit rate and distortion value.
6. processing unit in loop according to claim 5, wherein, described difference compensation arrangement comprises:
One judging unit, more described first bit rate and distortion value and described second bit rate and distortion value, when described first bit rate and distortion value are less than described second bit rate and distortion value, and described first bit rate and distortion value are when being less than zero, described first offset exports as an inconsistent signal by described judging unit; When described second bit rate and distortion value are less than described first bit rate and distortion value, and when described second bit rate and distortion value are less than zero, described second offset exports as described inconsistent signal by described judging unit.
7. processing unit in loop according to claim 1, wherein, described inconsistent signal comprises mode index, style index or binary numeral index and described first offset or described second offset.
8. processing unit in loop according to claim 1, wherein, the rear end that removes block effect filter is located at by described difference compensation arrangement, for going the output signal of block effect filter described in receiving.
9. an optionally pixel value difference compensation method, comprises:
Receive the rear reconstructed image data of a coding, after described coding, reconstructed image data includes multiple square;
Receive the front view data of a coding; And
According to the pixel of each described square difference status of the same position of reconstructed image data after view data and described coding before described coding, add up described difference status to produce multiple offset, optionally choose a described offset to produce an inconsistent signal.
10. method according to claim 9, wherein, described difference status includes lines difference and smoothness difference.
11. methods according to claim 10, wherein, described lines difference is the spike and low ebb that utilize multiple angle and edge to be formed and carries out classifying, count multiple offsets of each classification, a default equation is utilized to calculate bit rate and the distortion value of each described offset, select one first offset to produce described inconsistent signal according to described multiple bit rate and distortion value, and described bit rate corresponding to described first offset and distortion value are one first bit rate and distortion value.
12. methods according to claim 11, wherein, the Luminance Distribution of described square is divided into multiple section by described smoothness difference, count multiple offsets of each described section, a default equation is utilized to calculate bit rate and the distortion value of each described offset, select one second offset to produce described inconsistent signal according to described multiple bit rate and distortion value, and described bit rate corresponding to described second offset and distortion value are one second bit rate and distortion value.
13. methods according to claim 12, wherein, described method also comprises and described first bit rate and distortion value and described second bit rate and distortion value is compared, and determines which in described first offset and described second offset to export as described inconsistent signal according to comparative result.
14. methods according to claim 13, wherein, when described first bit rate and distortion value and described second bit rate and distortion value are less than zero, export described inconsistent signal.
15. methods according to claim 13, wherein, when described first bit rate and distortion value and described second bit rate and distortion value are more than or equal to zero, do not export described inconsistent signal.
16. 1 kinds of coded systems, comprise:
One motion compensating device, in order to compensate the displacement error of view data before reconstructed image data and current encoder after last coding; And a difference compensation arrangement, couple described motion compensating device, to receive after described current encoder view data before reconstructed image data and described current encoder, wherein, after described current encoder, the picture of reconstructed image data divides into multiple square, described difference compensation arrangement, according to the pixel of each described square multiple difference status of the same position of reconstructed image data after view data and described current encoder before described current encoder, produces an inconsistent signal;
Wherein, described difference compensation arrangement optionally selects the numerical value of the part variation state of described multiple difference status as offset, produces described inconsistent signal.
17. systems according to claim 16, wherein, described system also comprises a de-blocking filtering device, and in order to rebuild the block effect of image after current encoder described in filtering, and described difference compensation arrangement receives the output of described de-blocking filtering device.
18. systems according to claim 16, wherein, described difference compensation arrangement is coupled between a conversion quantization device and described motion compensating device.
19. systems according to claim 16, wherein, described system also comprises a random degree encoder, and described random degree encoder is according to reconstructed image data after current encoder described in described inconsistent signal compensation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310236938.9A CN104243996B (en) | 2013-06-14 | 2013-06-14 | The pixel value difference compensation method of processing unit and selectivity and coded system in loop |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310236938.9A CN104243996B (en) | 2013-06-14 | 2013-06-14 | The pixel value difference compensation method of processing unit and selectivity and coded system in loop |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104243996A true CN104243996A (en) | 2014-12-24 |
| CN104243996B CN104243996B (en) | 2017-07-28 |
Family
ID=52231213
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310236938.9A Active CN104243996B (en) | 2013-06-14 | 2013-06-14 | The pixel value difference compensation method of processing unit and selectivity and coded system in loop |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104243996B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110910805A (en) * | 2019-12-31 | 2020-03-24 | 深圳创维－Rgb电子有限公司 | Compensation adjusting device and compensation adjusting method of display screen |
| CN111654704A (en) * | 2020-05-26 | 2020-09-11 | 新相微电子(上海)有限公司 | Automatic multi-mode image compression control method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1154049A (en) * | 1995-12-29 | 1997-07-09 | 大宇电子株式会社 | Method for rectifying channel errors in image signal transmitted from block-based encoder |
| CN1728182A (en) * | 2004-06-29 | 2006-02-01 | 英特尔公司 | Image edge filtering |
| US20080298472A1 (en) * | 2007-06-04 | 2008-12-04 | Texas Instruments Incorporated | Throughput Performance When Applying Deblocking Filters On Reconstructed Image Frames |
-
2013
- 2013-06-14 CN CN201310236938.9A patent/CN104243996B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1154049A (en) * | 1995-12-29 | 1997-07-09 | 大宇电子株式会社 | Method for rectifying channel errors in image signal transmitted from block-based encoder |
| CN1728182A (en) * | 2004-06-29 | 2006-02-01 | 英特尔公司 | Image edge filtering |
| US20080298472A1 (en) * | 2007-06-04 | 2008-12-04 | Texas Instruments Incorporated | Throughput Performance When Applying Deblocking Filters On Reconstructed Image Frames |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110910805A (en) * | 2019-12-31 | 2020-03-24 | 深圳创维－Rgb电子有限公司 | Compensation adjusting device and compensation adjusting method of display screen |
| CN111654704A (en) * | 2020-05-26 | 2020-09-11 | 新相微电子(上海)有限公司 | Automatic multi-mode image compression control method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104243996B (en) | 2017-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6648315B2 (en) | Quantized block inverse quantization method | |
| CN1094291C (en) | Loop filter and loop filtering method | |
| KR101508292B1 (en) | Method And Apparatus For Intra Prediction | |
| KR100736104B1 (en) | Video coding method and apparatus supporting independent parsing | |
| US20100177819A1 (en) | Method and an apparatus for processing a video signal | |
| JP5838258B2 (en) | Method and apparatus for lossy compression encoding data and corresponding method and apparatus for reconstructing data | |
| CN102420988B (en) | Multi-view video coding system utilizing visual characteristics | |
| CN102939753A (en) | Method and apparatus for encoding video by compensating for pixel value according to pixel groups, and method and apparatus for decoding video by the same | |
| CN107071411A (en) | Intra-frame prediction method and the encoder and decoder using this method | |
| CN1145566A (en) | Post-processing device for eliminating blocking artifact and method thereof | |
| CN105100805A (en) | Method for decoding video signals | |
| CN105338352A (en) | Adaptive dequantization method and device in video decoding | |
| CN1998152A (en) | Conditional lapped transform | |
| CN108141593A (en) | For be directed to the efficient intraframe coding of deep video based on the discontinuous method of depth | |
| CN104883565A (en) | Decision-making method and device for intra-frame prediction mode of high efficiency video coding | |
| CN107864380A (en) | 3D HEVC fast intra-mode prediction decision-making techniques based on DCT | |
| WO2022078304A1 (en) | Video decoding method and apparatus, computer readable medium, program, and electronic device | |
| WO2022062880A1 (en) | Video decoding method and apparatus, computer readable medium, and electronic device | |
| CN105898300A (en) | Improved transformation coefficient sign data hiding method based on recovered transformation coefficient | |
| CN104243996A (en) | In-loop processor, selective pixel difference value compensation method and encoding system | |
| CN1825894A (en) | All phase cosine double orthogonal transformation and JPEG improving method | |
| CN105791868A (en) | Video coding method and equipment | |
| CN1317125A (en) | Method and device for estimating motion in digitized image with pixels | |
| CN113395516B (en) | Intra-frame prediction method and device, and computer readable storage medium | |
| Kutil | A generalization of quad-trees applied to image coding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |