WO2009002321A1 - Amélioration de qualité d'image - Google Patents

Amélioration de qualité d'image Download PDF

Info

Publication number
WO2009002321A1
WO2009002321A1 PCT/US2007/015114 US2007015114W WO2009002321A1 WO 2009002321 A1 WO2009002321 A1 WO 2009002321A1 US 2007015114 W US2007015114 W US 2007015114W WO 2009002321 A1 WO2009002321 A1 WO 2009002321A1
Authority
WO
WIPO (PCT)
Prior art keywords
stream
resolution
compressed code
compressed
enhancement
Prior art date
Application number
PCT/US2007/015114
Other languages
English (en)
Inventor
Rajan Laxman Joshi
James Arthur Fancher
Ana Belen Benitez
Original Assignee
Thomson Licensing
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
Application filed by Thomson Licensing filed Critical Thomson Licensing
Priority to PCT/US2007/015114 priority Critical patent/WO2009002321A1/fr
Publication of WO2009002321A1 publication Critical patent/WO2009002321A1/fr

Links

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
    • H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
    • H04N19/33—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability in the spatial domain
    • 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
    • H04N19/46—Embedding additional information in the video signal during the compression process
    • 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
    • H04N19/48—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using compressed domain processing techniques other than decoding, e.g. modification of transform coefficients, variable length coding [VLC] data or run-length data
    • 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
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/63—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets

Definitions

  • the present principles relate to image processing. More particularly, at least one implementation relates to resolution scalability for imaging systems.
  • Imaging system is used to encompass systems which process and render still images as well as moving images or videos. Many of the known imaging systems provide access to a video frame at different resolutions. For example, in 4K workflows for Digital Cinema (DC) or Digital Intermediates (DI), lower resolution versions (proxies) of the 4K resolution images are available.
  • the video frames are stored in a compressed format, and a lower resolution version of the video frame (in compressed format) can be obtained from the higher resolution compressed code-stream by simple truncation or parsing.
  • parsing is used in the sense of accessing non-contiguous parts of the compressed code-stream without performing decompression.
  • a system may decode the high resolution video frame and perform downscaling to the lower resolution. Downscaling allows system providers to use different algorithms for downscaling.
  • a compressed code-stream is created for an image.
  • the compressed code-stream is compliant with a DCI-like standard such that a higher- resolution compressed bitstream and a lower-resolution compressed bitstream can both be extracted from the compressed code-stream before decompressing the compressed code- stream.
  • An enhancement-layer compressed code-stream is created for enhancing an image corresponding to the lower-resolution compressed bitstream.
  • a data package is formed that includes the compressed code-stream and at least one of (1) the enhancement-layer compressed code-stream or (2) a pointer to the enhancement-layer compressed code-stream. The data package is sent for receipt by a device configured to extract the lower-resolution compressed bitstream, and to enhance the image corresponding to the lower-resolution compressed bitstream by using the enhancement-layer compressed code-stream.
  • a signal represents information in a data package.
  • the signal includes a first portion and a second portion.
  • the first portion represents a compressed code-stream for an image.
  • the compressed code-stream is compliant with a DCI-like standard such that a higher-resolution compressed bitstream and a lower-resolution compressed bitstream can both be extracted from the compressed code- stream before decompressing the compressed code-stream.
  • the second portion represents at least one of (1) an enhancement-layer compressed code-stream or (2) a pointer to the enhancement-layer compressed code-stream.
  • the enhancement-layer compressed code- stream is for enhancing an image corresponding to the lower-resolution compressed bitstream.
  • a data package is accessed.
  • the data package includes a compressed code-stream for an image and at least one of (1) an enhancement-layer compressed code-stream or (2) a pointer to the enhancement-layer compressed code-stream.
  • the compressed code-stream is compliant with a DCI-like standard such that a higher- resolution compressed bitstream' and a lower-resolution compressed bitstream can both be extracted from the compressed code-stream before decompressing the compressed code-
  • An enhanced lower-resolution image is generated based on the lower-resolution compressed bitstream and the enhancement-layer compressed code-stream.
  • _ _ _5l5 ⁇ r mat implementations may be configured or embodied in various manners. For example,0 an implementation may be performed as a method, or embodied as an apparatus configured to perform a set of operations or an apparatus storing instructions for performing a set of operations. Other aspects and features will become apparent from the following detailed description considered in conjunction with the accompanying drawings and the claims. 5 BRIEF DESCRIPTION OF THE DRAWINGS
  • Figure 1 is a flow diagram of a method for creating a low resolution enhancement layer to be used in conjunction with a resolution scalable encoding of a high resolution image.
  • Figure 2 is a flow diagram of a method for creating an enhanced low resolution image using the created low resolution enhancement layer of Figure 1.
  • Figure 3 is a flow diagram of a method for producing a 2K enhancement layer compressed code-stream.
  • Figure 4 is a flow diagram of a method for producing an enhanced 2K resolution image using the produced 2K enhancement layer compressed code-stream of Figure 3.
  • Figure 5 is a flow diagram of another implementation of a method for producing a 2K enhancement layer compressed code-stream.
  • Figure 6 is a flow diagram of another implementation of a method for producing an enhanced 2K resolution image using the produced 2K resolution enhancement layer compressed- code stream of Figure 5.
  • Figure 7 is .a block diagram of an encoder for producing a low resolution enhancement layer.
  • Figure 8 is a block diagram of a decoder for producing an enhanced low resolution image.
  • Figure 9 is a block diagram of an encoder for producing a 2K enhancement layer compressed code-stream according to another implementation.
  • Figure 10 is a block diagram of a decoder for producing an enhanced 2K resolution image according to another implementation.
  • Figure 11 is a flow diagram of an encoding method for producing an enhanced low resolution image according to an implementation.
  • Figure 12 is a flow diagram of a decoding method for producing an enhanced low resolution image according to an implementation.
  • Figure 13 is a flow diagram of a DCP-Specifi cation-compliant method for producing an enhanced 2K resolution image.
  • Figure 14 is a flow diagram of a DCP-Specification-compliant method for producing an enhanced 2K resolution image.
  • the compression method and file format used may place restrictions Qn the lower resolution image.
  • DCI Digital Cinema Initiative
  • Vl.0 July 2005
  • Digital Cinema Initiatives, LLC mandates that a 2K version should be contained in a 4K version and be accessible by simple truncation.
  • the downscaling filter is restricted to be the 9-tap low- pass filter used by JPEG2000.
  • a 2K projector has to use the 2K. extracted code-stream to reconstruct the 2K resolution video, it may be placed at a disadvantage compared to 2K projectors which have access to 2K compressed material which was produced by first downscaling the 4K resolution video and then compressing the downscaled 2K version.
  • performing the downscaling operation is typically computationally complex. It would be desirable to provide a method to enhance the quality of a low resolution image extracted from a scalable high resolution code-stream.
  • Enhancement layers are commonly used to provide quality scalability.
  • An example of this is MPEG-4 Fine Grain Scalability in the MPEG-4 video standard.
  • the purpose of the enhancement layer is to produce a reconstructed image that is closer to the original image.
  • an enhancement layer is used to produce a reconstructed image that is closer to a downscaled version of the original image—and not necessarily closer to the original image itself.
  • the inventors have recognized that such a counter-intuitive operation provides for technical advantages such as improved quality within the context of, for example, a DCI-like system or standard.
  • a DCI-like standard may be defined as a standard that provides a compressed code-stream from which both a higher- resolution compressed bitstream and a lower-resolution compressed bitstream can be extracted before decompressing the compressed code-stream.
  • the enhancement layer implementation referred to above also uses different coding processes to produce the original encoding (layer) and the additional (enhancement) encoding (layer). This is different from systems that use a common algorithm to gradually create and send more granular information for the iterative encoding of a given image. In such systems, each of the successive encodings (layers) is related to each other. An example is a system that sends, in each layer, increasingly fine quantizations of DCT (discrete cosine transform) coefficients.
  • DCT discrete cosine transform
  • At least one implementation provides a method for increasing or enhancing the quality of a low resolution image extracted from the scalable high resolution code-stream.
  • the downscaling method of the present principles utilizes an enhancement layer to produce an enhanced low resolution image by directly operating on the high resolution image.
  • the high resolution image is a 4K (4096 * 2160) image and the low resolution image is a 2K image (2048 * 1080).
  • the high resolution image is a 4K (4096 * 2160) image and the low resolution image is a 2K image (2048 * 1080).
  • the 4K compressed coderStream is fed to a 2K decoder, the decoder extracts a compressed code-stream corresponding to the 2K resolution and then decodes it to produce a 2K resolution extracted image.
  • this 2K resolution image is constrained to be the compressed version of the LL band (as known, the LL band is low frequency in both the horizontal and vertical directions) at the next lower resolution from the resolution hierarchy produced by the wavelet transform used in JPEG2000.
  • this may not be the best choice for a downscaling operation.
  • a downscaling method may be able to use a much better filter and adaptive processing to produce an improved or enhanced 2K image.
  • the projector is a 2K projector having a 2K decoder, it may be better to perform the downscaling operation on the higher resolution 4K original image to produce the downscaled 2K image.
  • the 2K image can be compressed at the maximum rate allowed by the DCI specification to produce a 2K distribution which is fed to the 2K projector.
  • the single inventory being the 4K distribution. This compares to having to provide in inventory the 4K distribution as well as the improved or enhanced 2K image.
  • the 2K projectors may be disadvantaged.
  • Figure 1 shows one implementation of creating a low resolution enhancement layer to be used in conjunction with a resolution scalable encoding of a high resolution image to produce an enhanced low resolution image.
  • a high resolution original image 100 (i.e., 4K image) is accessed and undergoes resolution scalable, encoding 102 to produce a high resolution compressed code-stream 114.
  • Accessing is a broad term, including, for example, obtaining, retrieving, receiving, manipulating, or processing in various manners.
  • encoding and “compression”, are used herein interchangeably.
  • the DCI Spec (JPEG2000) encoding essentially comprises DC IeVeI- "** " shifting, inter-component transform, wavelet transform, quantization, and entropy coding.
  • the high resolution original image 100 also undergoes downscaling 104 to produce a low resolution downscaled image 105.
  • the downscaling could be any desired downscaling operation, including a proprietary operation, to produce a downscaled image having a desired quality or a desired characteristic.
  • a low resolution compressed code-stream 107 is extracted (106) from the high resolution compressed code-stream 114, and the low resolution compressed code-stream 107 undergoes low resolution decoding 108 to produce a low resolution extracted image 109. Taking the difference (110) between the low resolution downscaled image 105 and the low resolution extracted image 109 produces a low resolution enhancement layer 111.
  • the enhancement layer 111 is encoded 112 to produce a low resolution enhancement layer compressed code-stream 116.
  • the implementation produces additional information 111.
  • This additional information 111 is, in a particular implementation of Figure 1, referred to as the enhancement layer 111.
  • the additional information is exactly the same as an enhancement layer.
  • the additional information is not referred to as an enhancement layer until after the additional information is formatted into a particular format of an enhancement layer.
  • the size of the enhancement layer compressed code-stream 116 may not exceed a fixed number of bytes. The byte limit may be determined, for example, based on user input.
  • Figure 2 shows how a low resolution enhancement layer compressed code-stream 116 can be combined with a corresponding high resolution compressed code-stream 114 to produce an enhanced low resolution image.
  • a low resolution compressed code-stream 204 is extracted 202 from the high resolution compressed code-stream 114.
  • the low resolution compressed code-stream 204 is then low resolution decoded 206 to obtain a low resolution extracted image 208 (expected to be the same as image 109).
  • a low resolution enhancement layer 212 is produced by decoding (210) the low resolution enhancement layer compressed code-stream 116.
  • the decoded low resolution enhancement layer 212 is added 214 to the low resolution extracted image 208 to produce an enhanced low resolution image 216.
  • the enhanced low resolution image 216 is expected to be the same as the low resolution downscaled image 105 if lossless encoding is used to encode the low resolution enhancement layer (112).
  • Those skilled in the art will also recognize that it is possible to create multiple low resolution enhancement layers, each corresponding to a different resolution. As an example, you could have a 2K enhancement layer and another IK enhancement layer. Each layer, when combined with the corresponding extracted resolution would create an enhanced image at that lower resolution.
  • FIGS 7 and 8 show block diagrams of an encoder 700 and decoder 800 according to the implementation shown in Figures 1 and 2, respectively.
  • the encoder 700 includes a processor 702 configured to receive high resolution original image data, a memory 703 in communication with the processor and a high resolution processing group 720 and low resolution processing group 730.
  • the processor 702 is directly coupled to the memory 703, the encoder 704, and the downscaler 706.
  • the processor 702 is indirectly coupled to the extractor 708, the decoder 709, and other units within the encoder 700.
  • the term “coupled” refers to both direct (no intervening units) and indirect (intervening units) connections, and such connections may be, for example, wired or wireless, and permanent or transient.
  • the processor 702 sends the high resolution image data to both the high resolution processing group 720 and the low resolution processing group 730.
  • the high resolution processing group 720 inputs the high resolution original image data to a resolution scalable encoder 704.
  • the encoder 704 outputs the high resolution compressed code-stream which is both input to the extractor 708 and to the transmitter 712.
  • the extractor 708 extracts a low resolution compressed code-stream from the high resolution compressed code-stream, and inputs the same to a low resolution decoder 709 to produce a low resolution extracted image.
  • the low resolution processing group 730 inputs the high resolution original image data to a downscaler 706 which downscales the high resolution original image data to a low resolution downscaled image.
  • the subtracter 710 outputs the difference between the low resolution downscaled image and the low resolution extracted image from the high resolution processing group 720 to produce a low resolution enhancement layer.
  • the low resolution enhancement layer is encoded 711 to produce a low resolution enhancement layer compressed code-stream, which is input to the transmitter 712 for transmission.
  • implementation of the transmit/send step 712 may include, for example, storing the high resolution compressed code-stream and the enhancement layer compressed code-stream on a hard drive or other physical media and transporting it to another location.
  • the enhancement layer is generated by the combined action of both the high resolution processing group 720 and the low resolution processing group 730. Accordingly, the combination-of the two groups 720 and 730 is referred to as an enhancement layer generator.
  • Figure 8 shows a decoder side 800 according to an implementation.
  • the data received from the encoder 700 which includes both the high resolution compressed code- stream and the low resolution enhancement layer compressed code stream, is received by a processor 802 or other processing device which may include a memory 803.
  • the processor 802 directs the high resolution compressed code stream to the high resolution processing group 820 and directs the low resolution enhancement layer compressed code stream to the low resolution processing group 830.
  • the high resolution processing group 820 inputs the high resolution compressed code stream to an extractor 804 to extract a low resolution compressed code stream.
  • the low resolution compressed code stream output from the extractor 804 is low resolution decoded 806 to produce a low resolution extracted image.
  • the low resolution processing group 830 directs the low resolution enhancement layer compressed code stream to an enhancement layer decoder 808.
  • the decoded low resolution enhancement layer is added 810 to the low resolution extracted image (generated from the high resolution image data) to produce an enhanced low resolution image.
  • the image could then be, for example, used for display on a
  • the enhanced image is generated by the combined action of both the high resolution processing group 820 and the low resolution processing group 830. Accordingly, the combination of the two groups 820 and 830 is referred to as an enhanced image generator.
  • the above implementation of Figures 2-3 and 7-8 illustrate a spatial-domain enhancement layer.
  • the enhancement layer includes data determined by, for example, subtracting (110) two images (109, 105) fromreach other. The images are all in the spatial domain, and typically include pixel values.
  • the enhancement layer includes data determined by, for example, subtracting (322) two sets (314, 320) of wavelet coefficients from each other. The coefficients are part of frequency- domain representations of the underlying images.
  • FIG. 3 A flowchart for producing a 2K enhancement layer compressed code-stream in accordance with an implementation is shown in Figure 3.
  • An original 4K image 300 is encoded 302 as per the DCI specification (JPEG2000) to produce a DCI compliant 4K compressed code-stream 306.
  • a 2K resolution compressed code-stream 310 is extracted 308 from the DCI compliant 4K code-stream 306 by simple truncation.
  • the 2K resolution extracted compressed code-stream 310 undergoes entropy decoding and dequantization 312 to produce reconstructed wavelet coefficients 314 for the extracted 2K image.
  • Operation 312 does not include, in this implementation, inverse wavelet transformation because the wavelet coefficients, rather than for example the spatial- domain pixel data, are desired.
  • the original 4K image undergoes downscaling by a factor of 2 in the horizontal and vertical direction 304 to produce a downscaled 2K image 316.
  • the downscaled 2K image 316 is DC level-shifted (318) to produce pixel values centered around zero; and then undergoes irreversible color transform (ICT) (318) as specified in the JPEG2000 standard, and wavelet transformation (318) with the JPEG2000 (9,7) filters to produce wavelet coefficients 320 for the downscaled 2K image.
  • ICT irreversible color transform
  • the ICT (34-8) is applied only if the DCI 4K encoding step (302) uses ICT.
  • the number of decomposition levels for the wavelet transformation is one less than the decomposition levels used for the DCI 4K encoding.
  • the reconstructed wavelet coefficients 314 for the extracted 2K image are subtracted 322 from the wavelet coefficients 320 for the downscaled 2K image to produce wavelet coefficients 324 for the 2K enhancement layer.
  • the wavelet coefficients 324 for the 2K enhancement layer are quantized and entropy encoded using JPEG2000 (326) to produce a 2K enhancement layer compressed code-stream 328.
  • the quantization level can be set to produce, at most, a fixed number of bytes, which can be user-specified.
  • Figure 4 shows how a 2K enhancement layer compressed code-stream can be combined with a corresponding DCI-compliant 4K compressed code-stream to produce the enhanced 2K image.
  • a 2K resolution compressed code-stream 404 is extracted 402 from a DCI-compliant 4K compressed code-stream 306.
  • the 2K resolution compressed code- stream 404 undergoes entropy decoding and dequantization 406 in the JPEG2000 framework to produce reconstructed wavelet coefficients for the extracted 2K image 408.
  • a corresponding 2K enhancement layer compressed code-stream 328 undergoes entropy decoding and dequantization 410 in the JPEG2000 framework to produce reconstructed wavelet coefficients for the 2K enhancement layer 412.
  • the reconstructed wavelet coefficients for the extracted 2K image 408 and for the 2K enhancement layer 412 are added 414 to produce the" wavelet coefficients for the enhanced 2K image 416.
  • the wavelet coefficients for the enhanced extracted 2K image 416 undergo wavelet synthesis and inverse ICT; and are then DC level-shifted and clipped 418 to the appropriate bit-depth to produce an enhanced 2K image 420.
  • the inverse ICT is applied only if the DCI-compliant 4K compressed code-stream 306 was produced using ICT.
  • the wavelet synthesis refers to a reverse wavelet transformation. -e-
  • Figures 5 and 6 show an alternative implementation for producing an enhanced 2k image from an original 4k image according to the present principles.
  • the implementation of Figures 5 and 6 is also DCI-compliant, but works with images data (spatial domain) rather than wavelet coefficients (frequency domain) or other descriptors of the image data.
  • an original 4K image 502 is DCI 4K encoded 504 as per the DCI specification to produce a DCI compliant 4K compressed code-stream 506.
  • a 2K resolution compressed code-stream 510 is extracted 508 from the DCI compliant 4K code- stream by simple truncation.
  • the 2K resolution extracted compressed code-stream 510 undergoes JPEG2000 decoding (including inverse wavelet transformation) 512 to produce a reconstructed extracted 2K image 514.
  • JPEG2000 decoding includes entropy decoding, dequantization, inverse ICT (if needed), DC level-shifting, and clipping to the appropriate bit-depth.
  • the original 4K image also undergoes downscaling by a factor of 2 in the horizontal and vertical direction (516) to produce a downscaled 2K image 518.
  • the reconstructed extracted 2K image 514 is subtracted 520 from the downscaled 2K image 518 to produce a 2K enhancement layer 522.
  • the 2K enhancement layer 522 is encoded 524 using JPEG2000 to produce a 2K enhancement layer compressed code-stream 526.
  • the quantization level can be set to produce at most a fixed number of bytes, which can be user-specified.
  • a 2K resolution compressed code-stream 604 is extracted from a DCI- compliant 4K compressed code-stream 506.
  • the 2K resolution compressed code-stream 604 undergoes JPEG2000 decoding 606 to produce a reconstructed extracted 2K image 608.
  • a corresponding 2K enhancement layer compressed code-stream 526 undergoes JPEG2000 decoding 610 to produce a reconstructed 2K enhancement layer 612.
  • the reconstructed extracted 2K image 608 and the reconstructed 2K enhancement layer 612 are added 614 and clipped 616 to the appropriate bit-depth, to produce an enhanced 2K image 618.
  • the enhanced 2K image 618 is expected to correspond to the downscaled 2K image 518 if lossless encoding is used to encode 524 the 2K enhancement layer 522.
  • Figure 9 shows an implementation of an encoder 900 that can be used to implement the method of Figure 3.
  • Figure 10 shows an implementation of a decoder 1000 that can be used to implement the method of Figure 4.
  • the original 4k image is received by a processor 902 or other input device.
  • the processor 902 is coupled to a memory 903 and is configured to input the original 4k image to a high resolution processing group 920 and a low resolution processing group 930.
  • the high resolution processing group 920 receives the original 4K image and inputs it to a DCI 4K encoder 904.
  • the DCI 4k encoder 904 produces a DCI compliant 4K compressed code stream, which is input to a transmitter 918 for transmission and is also input to an extractor 906 to extract a 2K resolution compressed code stream from the DCI compliant 4K compressed code stream.
  • the 2K resolution compressed code stream is decoded 908 (entropy decoding and dequantization) to produce reconstructed wavelet coefficients for the extracted 2K image.
  • the decoding in operation 908 may also be referred to as a partial decoding because the code stream is only decoded to the point of providing the wavelet coefficients and no inverse wavelet transformation is performed.
  • the operations of extraction (906) and decoding (908) can be generally referred to collectively as processing or even decoding.
  • the low resolution processing group 930 receives the original 4K image and downscales (910) the same to produce a downscaled 2K image. Wavelet coefficients for the — downscaled 2K image are generated by level shifting and wavelet transformation (912).
  • wave difference between reconstructed wavelet coefficients for the extracted 2K image (generated by the high resolution processing group) and the wavelet coefficients for the downscaled 2K image is determined by a subtracter 914 to produce wavelet coefficients for the 2K enhancement layer.
  • These wavelet coefficients are enhancement layer encoded (916) and sent to the transmitter 918 for transmission.
  • the transmit/ send unit 918 may combine the encoded enhancement layer (916) and the DCI compliant 4K compressed code stream (904) into a single transmission or storage unit.
  • unit 918 may form a DC package that includes that includes the compressed code-stream and at least one of (1) the enhancement-layer compressed code-stream or (2) a pointer to the enhancement-layer compressed code-stream. Accordingly, unit 918 may be referred to as a "combiner”.
  • the enhancement layer is generated by the combined action of both the high resolution processing group 920 and the low resolution processing group 930. Accordingly, the combination of the two groups 920 and 930 is referred to as an enhancement layer generator.
  • FIG. 10 shows a decoder 1000 according to an implementation.
  • the decoder 1000 includes a processor 1002 and memory 1003 where the processor is configured to receive the data from an encoder.
  • the data received from the encoder includes the DCI compliant 4K compressed code-stream and the 2K resolution enhancement layer compressed code-stream.
  • the processor 1002 sends the 4K resolution compressed code-stream to a high resolution processing group 1020 and the 2K resolution enhancement layer compressed code-stream to a low resolution processing group 1030.
  • the extractor 1004 extracts the 2K resolution compressed code-stream from the 4K resolution image data and the decoder (e.g., JPEG2000 decoder) 1006 decodes the extracted 2K resolution compressed code-sfream to produce reconstructed wavelet coefficients for the extracted 2K resolution image.
  • the low resolution processing group 1030 includes an enhancement layer decoder
  • the reconstructed wavelet coefficients for the extracted 2K resolution image (generated by the high resolution processing group) are then added (1010) with the reconstructed wavelet coefficients for the 2K resolution enhancement layer (generated by the low resolution processing group) to produce wavelet coefficients for the enhanced extracted low resolution image.
  • These wavelet coefficients are then synthesized, undergo inverse ICT, and are then level shifted and clipped (1012) to produce the enhanced 2K resolution image.
  • the inverse ICT step is performed only if the DCI compliant 4K compressed code-stream is generated using ICT. As before, these images may be displayed on a display 1014.
  • the enhanced image is generated by the combined action of both the high resolution processing group 1020 and the low resolution processing group 1030. Accordingly, the combination of the two groups 1020 and 1030 is referred to as an enhanced image generator.
  • Figures 9 and 10 may be adapted to implement the methods of Figures 5 and 6.
  • One of ordinary skill will readily understand how to make such modifications, particularly in light of the disclosure and discussion of Figures 7 and 8.
  • bit-depth for the 2K enhancement layer should be chosen so as to avoid unnecessary clipping.
  • the 4K and 2K images are 12-bit.
  • the 2K enhancement layer has to- be encoded as a signed 13 -bit image. If the subtraction is going to take place in the wavelet domain, as shown in Figure 3, then the analysis of the bit-depth needed may be more complex.
  • both the analysis filters (low pass and high pass) are implemented with (1,1) normalization as described by Taubman and Marcellin ("JPEG2000 image compression fundamentals, standards and practice" Kluwer Academic Publishers, ISBN 0-7923-7519-x, 2002).
  • N bit-depth of the original 4K image
  • (N+ 1) bits are sufficient to represent each subband without any clipping.
  • the downscaled 2K image also has a bit-depth of N
  • its wavelet coefficients can be represented by (N+ 1) bits.
  • the wavelet coefficients for the 2K enhancement layer which are calculated as a difference, can be represented by (N+2) bits.
  • the wavelet coefficients of the enhancement layer have been generated by wavelet decomposition of a (N+ 1) bit signed image.
  • the 2k enhancement layer may be compressed using any other compression method instead of JPEG2000 encoding.
  • the wavelet coefficients for the 2K enhancement layer may be compressed using any other compression method instead of JPEG2000 encoding.
  • JPEG2000 encoding techniques can be used for compressing the 2K enhancement layer without producing a JPEG2000 compliant code-stream. This is accomplished, for example, by forcing the 2K enhancement layer encoding engine to use the equivalent parameters as the 4K JPEG2000 encoding engine. In that case, it is not necessary to store a number of marker segments. But the resulting gain in compression efficiency is typically small.
  • a DC composition is defined as an ordered sequence of reels. Each reel can contain multiple media track files to be reproduced in parallel during the presentation. A reel can have a single main picture, main sound, and main subtitle track files. However, the current standard specification also supports additional track files for future uses.
  • a DC composition can be put into one or more DC packages for distribution to theaters or other venues.
  • the low resolution enhancement can be performed by maintaining the enhancement data separate from the DCP Stream.
  • Figures 11 and 12 show an exemplary implementation of this concept from the transmitter (i.e., creation) and receiving (i.e., playback) sides, respectively.
  • a DCI compliant 4K compressed code-stream 1102 is used to create DC picture track files 1106 and to create the 2K enhancement layer compressed code stream 1104.
  • the 2K enhancement layer compressed code stream 1104 is placed in a separate file 1108.
  • One or more DC packages are created 1110 using the previously created DC picture track files and which include the filename or a pointer to the enhancement layer file.
  • pointer is a broad term that includes, for example, an address, a name (e.g. a filename), or other descriptor that indicates where the enhancement layer file is located.
  • the filename, or pointer is included in a metadata field that the standard defines.
  • the metadata field may be, for example, a free-form text field, effectively allowing an implementation to use the field for a variety of purposes.
  • the one or more DC packages and the enhancement layer file can be sent 1112 (either manually via mail or electronically via communication network of any suitable type).
  • the enhancement layer file can be transmitted completely separate from the DC package information.
  • the creation of the DCI compliant 4K compressed code stream and the 2K enhancement layer compressed code-stream may be, for example, as described above. "**
  • Figure 12 shows the method 1200 for receiving and processing the enhancement file data in accordance with an implementation.
  • one or more DC packages are received 1202.
  • a determination 1204 is then made as to whether the DC packages include a filename or other pointer to an enhancement layer file. If there is no enhancement layer filename or pointer, the DC picture track is decoded and the low resolution (2K) image is generated in accordance with the DCI spec 1206. If there is an enhancement filename or pointer contained within the DC package information, the enhancement layer file named in the metadata is identified, located, and accessed 1208. Once identified and accessed, the DC picture track file(s) is decoded and the enhanced low resolution (2K) image is generated 1210 using the enhancement layer file data.
  • Figures 13 and 14 show a representative implementation of a DCP Specification compliant implementation of the present principles.
  • the method 1300 of Figure 13 shows that both the high resolution compressed code-stream and the low resolution enhancement layer compressed code-Stream are combined such that the low resolution enhancement layer compressed code-stream is included in the created DC packages. That is, the low resolution enhancement layer ' compressed code-stream is included in an additional picture track within the DC packages.
  • the high resolution compressed code-stream is included in the DC main picture track (1302) and the low resolution enhancement layer compressed code-stream is included into DC additional picture track file(s) (1304). Once included, the DC package(s) are created 1306.
  • the created DC packages can then be sent (i.e., manually in reels or other storage media) or transmitted electronically 1308 to the respective receivers (e.g., movie theaters) of the image data.
  • Figure 14 shows the receiving end 1400 of the DC package data.
  • the DC package data is received 1402, and a determination is then made (1404) as to whether the DC package data includes the low resolution enhancement layer data as an additional picture track file. If it does, the DC main and additional picture tracks are decoded (1408) and the enhanced low resolution image is generated using the extracted 2K image and the enhancement layer data contained in the DC main and additional picture track file (1408).
  • the DC package(s) are decoded (1406) and the low resolution (2K) image is generated in accordance with standard DCI Specification requirements.
  • the system typically ignores the additional picture track file.
  • a system may include an additional picture track file (in a DC package) for a first enhancement layer, as well as a pointer to a second enhancement layer.
  • a system may include in any given DC package only one of an additional pictur ⁇ track file or a pointer, but the system may allow a user to choose either mechanism for each DC package. Such a choice may be made, for example, in real time -based on the size of the enhancement layer for the given DC package.
  • a system could make the choice without user-intervention by, for example, consulting a look-up table that indicates the preferred mechanism for a given movie or a given intended-recipient of the DC package.
  • Figure 4 describes a decoder that requires at least one inverse wavelet transformation (unit 418), and Figure 6 describes a decoder that requires at least two inverse wavelet transformations (units 606 and 610).
  • Applications include, for example, a pre-processor or an encoder for creating DC packages or other data packages, or a decoder or other video receiving/processing apparatus, such as described above.
  • various units may be integrated, such as, for example, a pre-processor and an encoder.
  • the features and aspects herein described may, however, be adapted for other application areas.
  • the implementations described herein may be implemented in, for example, a method or process, an apparatus, or a software program. Even if only discussed in the context of a single form of implementation (for example, discussed only as a method), the implementation of features discussed may also be implemented in other forms (for example, an apparatus or program).
  • An apparatus may be implemented in, for example, appropriate hardware, software, and firmware.
  • the methods may be implemented in, for example, an apparatus such as, for example, a processor, which refers to processing devices in general, including, for example, a computer,.a microprocessor, an integrated circuit, or a programmable logic device.
  • Implementations of the various processes and features described herein may be embodied in a variety of different equipment or applications, particularly, for example, equipment or applications associated with video transmission.
  • equipment include video coders, video decoders, video codecs, web servers, and personal computers.
  • encodings may be sent or received over a variety of paths, including, for example, wireless or wired paths, the Internet, cable television lines, telephone lines, and Ethernet connections. Additionally, as should be clear, the equipment may be mobile and even installed in a mobile vehicle.
  • the methods may be implemented by instructions being performed by a processor, and such instructions may be stored on a processor readable medium such as, for example, an integrated circuit, a software carrier, or other storage device such as, for example, a hard disk, a compact diskette, a random access memory (“RAM”), or a read-only memory (“ROM").
  • a processor may also include a processor readable medium having, for example, instructions for carrying out a process.
  • implementations may also produce a signal formatted to carry information that may be, for example, stored or transmitted.
  • the signal may be transmitted as, for example, an electromagnetic wave, and may carry information by, for example, modulating one or more carrier frequencies.
  • the information may include, for example, instructions for performing a method, or data produced by one of the described implementations.

Landscapes

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

Abstract

On obtient avec une qualité améliorée une image basse résolution extraite d'un flux de bits de haute résolution échelonnable dans un flux de bits DCP. Un processus consiste à créer (302, 504) un flux de codes compressé pour une image de sorte qu'un flux de bits compressé de résolution plus élevée et qu'un flux de bits compressé de résolution moins élevée puissent être extraits avant la décompression. Un flux de codes compressé de couche d'amélioration (couche d'amélioration) est créé (524) pour améliorer une image correspondant au flux de bits de résolution moins élevée. Un paquet de données est formé (918, 1110, 1306) comprenant le flux de codes compressé et au moins la couche d'amélioration (1) ou un pointeur (2) vers la couche d'amélioration. Le paquet de données est envoyé (918, 1112, 1308) en vue d'une réception par un dispositif configuré pour extraire le flux de bits de résolution moins élevée et pour améliorer l'image correspondant au flux de bits de résolution moins élevée au moyen de la couche d'amélioration. L'invention concerne aussi un signal complémentaire et un processus de décodage, ainsi que des structures permettant d'effectuer ces processus.
PCT/US2007/015114 2007-06-27 2007-06-27 Amélioration de qualité d'image WO2009002321A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/US2007/015114 WO2009002321A1 (fr) 2007-06-27 2007-06-27 Amélioration de qualité d'image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/US2007/015114 WO2009002321A1 (fr) 2007-06-27 2007-06-27 Amélioration de qualité d'image

Publications (1)

Publication Number Publication Date
WO2009002321A1 true WO2009002321A1 (fr) 2008-12-31

Family

ID=39310342

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2007/015114 WO2009002321A1 (fr) 2007-06-27 2007-06-27 Amélioration de qualité d'image

Country Status (1)

Country Link
WO (1) WO2009002321A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011049946A1 (fr) 2009-10-19 2011-04-28 Synta Pharmaceuticals Corp. Combinaison de thérapie contre le cancer avec des composés inhibiteurs de hsp90
US8872981B1 (en) 2011-12-15 2014-10-28 Dolby Laboratories Licensing Corporation Backwards-compatible delivery of digital cinema content with extended dynamic range
US8891863B2 (en) 2011-06-13 2014-11-18 Dolby Laboratories Licensing Corporation High dynamic range, backwards-compatible, digital cinema

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5768535A (en) * 1995-04-18 1998-06-16 Sun Microsystems, Inc. Software-based encoder for a software-implemented end-to-end scalable video delivery system
WO2001047277A1 (fr) * 1999-12-20 2001-06-28 Sarnoff Corporation Codage vidéo échelonnable
WO2006054507A1 (fr) * 2004-11-19 2006-05-26 Matsushita Electric Industrial Co., Ltd. Méthode de codage d’image animée et méthode de décodage d’image animée

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5768535A (en) * 1995-04-18 1998-06-16 Sun Microsystems, Inc. Software-based encoder for a software-implemented end-to-end scalable video delivery system
WO2001047277A1 (fr) * 1999-12-20 2001-06-28 Sarnoff Corporation Codage vidéo échelonnable
WO2006054507A1 (fr) * 2004-11-19 2006-05-26 Matsushita Electric Industrial Co., Ltd. Méthode de codage d’image animée et méthode de décodage d’image animée
EP1816870A1 (fr) * 2004-11-19 2007-08-08 Matsushita Electric Industrial Co., Ltd. Méthode de codage video et méthode de décodage video

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
MARCELLIN M W ET AL: "JPEG2000 FOR DIGITAL CINEMA", SMPTE MOTION IMAGING JOURNAL, SMPTE INC., WHITE PLAINS, NY, US, vol. 114, no. 5/6, May 2005 (2005-05-01), pages 202 - 209, XP001229569, ISSN: 0036-1682 *
OHM J-R ET AL: "Interframe wavelet coding-motion picture representation for universal scalability", SIGNAL PROCESSING. IMAGE COMMUNICATION, ELSEVIER SCIENCE PUBLISHERS, AMSTERDAM, NL, vol. 19, no. 9, October 2004 (2004-10-01), pages 877 - 908, XP004607152, ISSN: 0923-5965 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011049946A1 (fr) 2009-10-19 2011-04-28 Synta Pharmaceuticals Corp. Combinaison de thérapie contre le cancer avec des composés inhibiteurs de hsp90
US8891863B2 (en) 2011-06-13 2014-11-18 Dolby Laboratories Licensing Corporation High dynamic range, backwards-compatible, digital cinema
US9781417B2 (en) 2011-06-13 2017-10-03 Dolby Laboratories Licensing Corporation High dynamic range, backwards-compatible, digital cinema
US8872981B1 (en) 2011-12-15 2014-10-28 Dolby Laboratories Licensing Corporation Backwards-compatible delivery of digital cinema content with extended dynamic range
US8922720B1 (en) 2011-12-15 2014-12-30 Dolby Laboratories Licensing Corporation Backwards-compatible delivery of digital cinema content with extended dynamic range

Similar Documents

Publication Publication Date Title
US8644632B2 (en) Enhancing image quality
Taubman et al. JPEG2000: Standard for interactive imaging
US8896717B2 (en) Methods for deploying video monitoring applications and services across heterogeneous networks
KR100681168B1 (ko) 미세 입상 스케일 가능한 비디오를 위한 잔류 신호의인코딩 및 디코딩 시스템 및 그 방법
KR101751221B1 (ko) 하위 호환성 ｖｄｒ 코덱을 위한 효율적인 트랜스코딩
JP2003503975A (ja) スケーラブルビデオコーディングのシステム及び方法
MX2013007030A (es) Grabador de svc-a-avc con multiplexor estadistico de circuito abierto.
JP2004254133A (ja) 動画再生システム、動画再生装置、動画送信装置、動画再生方法、プログラム、及び、記録媒体
US8243798B2 (en) Methods and apparatus for scalable video bitstreams
KR20070085316A (ko) 모바일 이미징 애플리케이션, 장치구조, 및 서비스 플랫폼구조
CN101888553B (zh) 用于可伸缩视频编码的方法和装置
WO2009002321A1 (fr) Amélioration de qualité d'image
US20140368672A1 (en) Methods for Deploying Video Monitoring Applications and Services Across Heterogeneous Networks
KR20070085317A (ko) 비디오 모니터링 애플리케이션, 장치구조, 및 시스템 구조
Auli-Llinas et al. Enhanced JPEG2000 quality scalability through block-wise layer truncation
Kobayashi et al. Extension of JPEG XS for two-layer lossless coding
US6345120B1 (en) Image processing system, image data transmission and reception apparatus, and image processing method
Skodras The JPEG2000 image compression standard in mobile health
JP4385928B2 (ja) スケーラブル動画像符号化・復号化方法及びシステム
JP3617358B2 (ja) 映像符号化方法および装置
Huang et al. An experimental analysis of DCT-based approaches for fine-grain multiresolution video
Qiao et al. Motion-JPEG2000 Stream Scaling for Multi-Resolution Video Transmission
Baccaglini et al. Robust Distributed Storage of Digital Cinema Contents
Smith et al. Simple video format for mobile applications
Auli-Llinas et al. Research Article Enhanced JPEG2000 Quality Scalability through Block-Wise Layer Truncation

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: 07835924

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 07835924

Country of ref document: EP

Kind code of ref document: A1