(19)
(11)EP 3 108 658 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
06.05.2020 Bulletin 2020/19

(21)Application number: 15762065.9

(22)Date of filing:  12.03.2015
(51)Int. Cl.: 
H04N 19/463  (2014.01)
H04N 19/593  (2014.01)
(86)International application number:
PCT/CN2015/074088
(87)International publication number:
WO 2015/135484 (17.09.2015 Gazette  2015/37)

(54)

METHOD OF COLOR INDEX CODING WITH PALETTE STUFFING

VERFAHREN ZUR FARBINDEXCODIERUNG MIT PALETTENFÜLLUNG

PROCÉDÉ DE CODAGE D'INDICE DE COULEUR AVEC BOURRAGE DE PALETTE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 14.03.2014 US 201461952917 P
31.03.2014 US 201461972584 P

(43)Date of publication of application:
28.12.2016 Bulletin 2016/52

(73)Proprietor: HFI Innovation Inc.
Hsinchu County (TW)

(72)Inventors:
  • SUN, Yu-Chen
    Keelung City 204 (TW)
  • CHUANG, Tzu-Der
    Zhubei City Hsinchu County 302 (TW)
  • CHEN, Yi-Wen
    Taichung City 420 (TW)
  • HUANG, Yu-Wen
    Taipei City 104 (TW)

(74)Representative: Hoefer & Partner Patentanwälte mbB 
Pilgersheimer Straße 20
81543 München
81543 München (DE)


(56)References cited: : 
EP-A2- 3 080 990
CN-A- 101 141 637
US-A1- 2011 110 416
CN-A- 1 925 612
US-A1- 2003 169 932
  
  • LAROCHE G ET AL: "Non-RCE4: Palette Prediction for Palette mode", 16. JCT-VC MEETING; 9-1-2014 - 17-1-2014; SAN JOSE; (JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11 AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-P0114-v3, 10 January 2014 (2014-01-10), XP030115611,
  • T-D CHUANG ET AL: "Non-RCE4: Major color table propagation through non-palette CUs", 16. JCT-VC MEETING; 9-1-2014 - 17-1-2014; SAN JOSE; (JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11 AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-P0096-v3, 9 January 2014 (2014-01-09), XP030115581,
  • ZHU W ET AL: "Non-RCE3:Template-based palette prediction", 14. JCT-VC MEETING; 25-7-2013 - 2-8-2013; VIENNA; (JOINT COLLABORATIVE TEAM ON VIDEO CODING OF ISO/IEC JTC1/SC29/WG11 AND ITU-T SG.16 ); URL: HTTP://WFTP3.ITU.INT/AV-ARCH/JCTVC-SITE/,, no. JCTVC-N0169-v3, 30 July 2013 (2013-07-30), XP030114648,
  • SUN, YU-CHEN ET AL.: 'SCCE3 Test A.2: Palette coding and palette predictor updates' JOINT COLLABORATIVE TEAM ON VIDEO CODING (JCT-VC) OF ITU-T SG 16 WP 3 AND ISO/IEC JTC 1/SC 29/WG 11 09 July 2014, pages 1 - 3, XP030116383
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

CROSS REFERENCE TO RELATED APPLICATIONS



[0001] The present invention claims priority to U.S. Provisional Patent Application, Serial No. 61/952,917, filed on March 14, 2014 and U.S. Provisional Patent Application, Serial No. 61/972,584, filed on March 31, 2014.

FIELD OF THE INVENTION



[0002] The present invention relates to color index coding for video data. In particular, the present invention relates to techniques to improve the performance of color index coding by using palette generated with palette stuffing. More particularly, the present invention relates to a method according to the pre-characterizing part of claim 1. Such a method is shown in document WO 2015/086718 A2. This document represents prior art according to Article 54(3) EPC.

BACKGROUND AND RELATED ART



[0003] High Efficiency Video Coding (HEVC) is a new coding standard that has been developed in recent years. In the High Efficiency Video Coding (HEVC) system, the fixed-size macroblock of H.264/AVC is replaced by a flexible block, named coding unit (CU). Pixels in the CU share the same coding parameters to improve coding efficiency. A CU may begin with a largest CU (LCU), which is also referred as coded tree unit (CTU) in HEVC. In addition to the concept of coding unit, the concept of prediction unit (PU) is also introduced in HEVC. Once the splitting of CU hierarchical tree is done, each leaf CU is further split into one or more prediction units (PUs) according to prediction type and PU partition.

[0004] Along with the High Efficiency Video Coding (HEVC) standard development, the development of extensions of HEVC has also started. The HEVC extensions include range extensions (RExt) which target at non-4:2:0 color formats, such as 4:2:2 and 4:4:4, and higher bit-depths video such as 12, 14 and 16 bits per sample. One of the likely applications utilizing RExt is screen sharing, over wired- or wireless-connection. Due to specific characteristics of screen contents, coding tools have been developed and demonstrate significant gains in coding efficiency. Among them, the color index coding (a.k.a. major color based coding) techniques represent block of pixels using indices to the palette (major colors), and encode the palette and the indices by exploiting spatial redundancy. While the total number of possible color combinations is huge, the number of colors in an area of picture is usually very limited for typical screen contents. Therefore, the color index coding becomes very effective for screen content materials.

[0005] During the early development of HEVC range extensions (RExt), several proposals have been disclosed to address palette-based coding. For example, a palette prediction and sharing technique is disclosed in JCTVC-N0247 (Guo et al.,"RCE3: Results of Test 3.1 on Palette Mode for Screen Content Coding", Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 14th Meeting: Vienna, AT, 25 July - 2 Aug. 2013 Document: JCTVC-N0247) and JCTVC-O0218 (Guo et al., "Evaluation of Palette Mode Coding on HM-12.0+RExt-4.1", Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 15th Meeting: Geneva, CH, 23 Oct. - 1 Nov. 2013, Document: JCTVC-O0218). In JCTVC-N0247 and JCTVC-O0218, the palette of each color component is constructed and transmitted. The palette can be predicted (or shared) from its left neighboring CU to reduce the bitrate. All pixels within the given block are then coded using their palette indices. An example of encoding process according to JCTVC-N0247 is shown as follows.
  1. 1. Transmission of the palette: the color index table size is first transmitted followed by the palette elements.
  2. 2. Transmission of pixel values: the pixels in the CU are encoded in a raster scan order. For each group of one or more pixels, a flag for a run-based mode is first transmitted to indicate whether the "run" mode or "copy-above" mode is being used.

    2.1 "run" mode: In the "run" mode, a palette index is first signaled followed by "palette_run" (e.g., M) representing the run value. The run value indicates that a total of M samples are all coded using "run" mode. No further information needs to be transmitted for the current position and the following M positions since they have the same palette index as that signaled in the bitstream. The palette index (e.g., i) may also be shared by all three color components, which means that the reconstructed pixel values are (Y, U, V) = (paletteY[i], paletteU[i], paletteV[i]) for the case of YUV color space.

    2.2 "copy-above" mode: In the "copy-above" mode, a value "copy_run" (e.g., N) is transmitted to indicate that for the following N positions (including the current one), the palette index is the same as the corresponding palette index in the row above.

  3. 3. Transmission of residue: the palette indices transmitted in Stage 2 are converted back to pixel values and used as the prediction. Residue information is transmitted using HEVC residual coding and is added to the prediction for the reconstruction.


[0006] In JCTVC-N0247, the palette of each component are constructed and transmitted. The palette can be predicted (shared) from its left neighboring CU to reduce the bitrate. In JCTVC-O0218, each element in the palette is a triplet, representing a specific combination of the three color components. The predictive coding of palette across CU is removed.

[0007] Yet another major color-base coding (color index coding) method is disclosed by Guo et al. in JCTVC-O0182 (Guo et al., "AHG8: Major-color-based screen content coding", Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 15th Meeting: Geneva, CH, 23 Oct. - 1 Nov. 2013, Document: JCTVC-O0182). According to JCTVC-O0182, the palette of each component are constructed and transmitted. However, instead of predicting the entire palette from the left CU, individual entry in a palette can be predicted from the exact corresponding palette entry in the above CU or left CU.

[0008] According to JCTVC-O0182, a predictive coding method is applied on the indices for transmission of pixel values, where a pixel line can be predicted by different modes. In specific, three kinds of line modes are used for a pixel line, i.e. horizontal mode, vertical mode and normal mode. In horizontal mode, all the pixels in the same line have the same value. If the value is the same as the first pixel of the above pixel line, only line mode signaling bits are transmitted. Otherwise, the index value is also transmitted. In vertical mode, the current pixel line is the same with the above pixel line. Therefore, only line mode signaling bits are transmitted. In normal mode, pixels in a line are predicted individually. For each pixel, the left or above neighboring pixel is used as predictor, and the prediction symbol is transmitted to the decoder. Furthermore, pixels are classified into major color pixels and escape pixels. For major color pixels, the decoder reconstructs the pixel values using the major color indices and palette. For escape pixels, the encoder will further send the pixel values.

BRIEF SUMMARY OF THE INVENTION



[0009] The above-mentioned objectives are achieved by a method according to independent claim 1. The dependent claims define preferred embodiments thereof.

[0010] A method of coding a block of video data using palette coding for a video coding system is disclosed. According to the method, when a previous block is available, the system receives first information related to a previous palette set corresponding to previous palettes/ triplet palette, and second information related to a previous palette predictor set corresponding to previous palette predictors/ triplet-palette predictors associated with the previous block. The system then derives a current palette predictor set corresponding to current palette predictors/ triplet-palette predictor associated with the current block based on the previous palette set and the previous palette predictor set by including at least one unused color or triplet of the previous palette predictor set in the current palette predictor set when at least one unused color or triplet of the previous palette predictor set in the current palette predictor set exists. A current palette set corresponding to current palettes/triplet-palette associated with the current block is derived from the current palette predictor set. After the current palette set is derived, palette encoding or decoding is applied to the current block according to the current palette set.

[0011] One or more next updating signals can be signaled for deriving a next palette predictor set based on the current palette set and the current palette predictor set. A number of previous palette predictors/triplet-palette predictor in the previous palette set corresponds to M, and M sets of previous palette predictors/triplet-palette predictor in the previous palette set can be maintained at an encoder side or a decoder side, and M is an integer greater than zero. The previous block and the current block may correspond to coding units or prediction units. The current palette predictor set may be derived depending on the coding-unit size if the previous block and the current block correspond to the coding units, or depending on the prediction-unit size if the previous block and the current block correspond to the prediction units.

[0012] The derivation process for the current palette predictor set may comprise updating the previous palette predictor set according to one or more updating signals. The updating signals may include a last palette-size or a new palette-size, and the current palette set is derived from the current palette predictor set by copying a number of beginning entries of the current palette predictor set, and the number corresponds to the last palette-size or the new palette-size. In another example, the updating signals may include new palette-size for each of current palettes/triplet-palette associated with the current block, a number of modification flags corresponding to new palette-size to identify whether the corresponding entries of current previous palette predictors/triplet-palette predictor in the previous palette set corresponds to M, and M sets of previous palette predictors/triplet-palette predictor in the previous palette set can be maintained at an encoder side or a decoder side, and M is an integer greater than zero. The previous block and the current block may correspond to coding units or prediction units. The current palette predictor set may be derived depending on the coding-unit size if the previous block and the current block correspond to the coding units, or depending on the prediction-unit size if the previous block and the current block correspond to the prediction units.

[0013] The derivation process for the current palette predictor set may comprise updating the previous palette predictor set according to one or more updating signals. The updating signals may include a last palette-size or a new palette-size, and the current palette set is derived from the current palette predictor set by copying a number of beginning entries of the current palette predictor set, and the number corresponds to the last palette-size or the new palette-size. In another example, the updating signals may include new palette-size for each of current palettes/triplet-palette associated with the current block, a number of modification flags corresponding to new palette-size to identify whether the corresponding entries of current palettes/triplet-palette associated with the current block are updated, and a new value for each to-be-modified entry as indicated by the corresponding modification flag. One or more modification flags can be context adaptive coded using context-adaptive binary arithmetic coding (CABAC). Also, one or more modification flags can be context adaptive coded using a bypass mode of CABAC. Modification flags can be limited to the beginning NP entries of current palettes/triplet-palette associated with the current block, and remaining entries are forced to be modified, where NP is a positive integer.

[0014] If the palette size of previous palettes/triplet-palette is smaller than a maximum palette size, then non-reused entries of previous palette predictors/triplet-palette predictor can be copied to current entries of current palette predictors/triplet-palette predictor after the current block is encoded or decoded by the palette coding. If the palette size of previous palettes/triplet-palette is larger than a maximum palette predictor size, then one or more updating values associated with the previous palette predictor set are not used by the previous palette set, and the updating values can be derived or transmitted. The derivation process of the current palette predictor set may include checking and removing redundancy in the current palette predictor set. The process of checking and removing redundancy can be performed before, during or after one or more new entries are placed into the current palette predictor set.

[0015] If the current block is at a first block in a slice, tile, largest coding unit row, or largest coding unit, then the previous block is invalid. In this case, the current palette set is derived by an encoder without the previous palette set nor the previous palette predictor set.

BRIEF DESCRIPTION OF THE DRAWINGS



[0016] 

Fig. 1A illustrates an example of mapping color values to color indices on a component-wise basis, where each color component uses its own palette.

Fig. 1B illustrates an example of mapping color values to triplets based on color combinations, where each color components use a triplet palette.

Fig. 2 illustrates an example of palette derivation from palette predictor using a palette sharing mode.

Fig. 3 illustrates an example of palette derivation from palette predictor using a normal mode based on palette updating signals.

Fig. 4 illustrates an example of palette derivation from palette predictor using a normal mode based on palette updating signals, where the un-reused entries in the previous palette predictor are propagated into the current predictor.

Fig. 5A illustrates an example of current palette predictor derived from a last coded palette according an existing color index coding.

Fig. 5B illustrates an example of current palette predictor derived from a last coded palette and the previous palette predictor according an embodiment of the present invention.

Fig. 5C illustrates an example of current palette predictor derived from a last coded palette and the previous palette predictor with redundancy checking according an embodiment of the present invention.

Fig. 6 illustrates an exemplary flowchart of a system for color index coding using a palette predicted from a palette predictor including one or more entries form a previous palette predictor according to an embodiment of the present invention.


DETAILED DESCRIPTION OF THE INVENTION



[0017] In the present invention, various techniques to improve the performance of color index coding are disclosed. In particular, a new copy run mode that uses one or more previous decoded indices to identify index prediction candidates for a current index. In one embodiment of the present invention, a layered color representation is disclosed as follows.

Layered Color Representation



[0018] In the palette coding mode, pixels in each CU are classified into major colors and escape pixels. Major colors are the most representative pixel values in a CU like a palette, which cover the majority pixels of the CU. Escape pixels are the pixels that do not belong to the major colors.

[0019] Layered color representation (LC) method is disclosed herein for major color representation. Based on the LC representation, each pixel is represented by a triplet color index. And the triplet color index (also named triplet index) is predictively coded and transmitted to indicate whether the pixel is a major-color pixel or not. If the pixel is a major-color pixel, further information is transmitted to indicate the major color it belongs to.

[0020] As an example of layered major colors for the CU, a two layers representation is used. The first layer is to describe the most representative color values in each component, called major colors. Based on major colors, the second layer is to identify the most representative combination, called major (color) combination.

A. Examples of layered color representation



[0021] For each CU, an encoder first calculates the histogram of color values on a component-wise basis and chooses a set of representative major color values. These major colors are stored as component-wise palettes. The encoder further chooses a set of representative color combinations of the major color values based on component-wise palettes. The combinations are recorded in a triplet palette format.

[0022] The layered color representation can be applied to color pixels in any color format (e.g., YUV444, YUV420, YUV422, YUV400 or RGB444). The color combination of the second layer can be applied on subset of components. For example, the color combination may correspond to UV only in the YUV420 format.

[0023] For YUV444 and RGB444, the combination of second layer can be applied to all components as Fig.1A and Fig. 1B. As shown in Fig. 1A, each of the three color components has its own color palette. The color indices for the three color components are from 0 to Ni, where i = 1, 2 or 3. Fig. 1B illustrates the three color components jointly use triplet index, N4.

[0024] With palettes and a triplet palette, a pixel can be represented by a triplet index. A decoder first use the triplet index to find out three major color indexes for the three color components from the triplet palette. The major color indexes are used to identify color values from the palettes and to reconstruct the pixels.

[0025] For YUV420 and YUV422, the combination of second layer can be applied to the UV components. Therefore, one palette index is used for Y component and one triplet index for the UV components to represent a color pixel. While there are only two color components (i.e., U and V) in this case, the term "triplet index" is also used to refer to two-color index.

B. Palette and triplet palette coding



[0026] During decoding, a decoder maintains M sets of palette predictors and triplet palette predictors. M can be transmitted or derived. M can be as small as one. If M is larger than one, the encoder may further signal which predictor is used or the decoder can derive it.

[0027] At the decoder side, a decoder can derive which predictor is used according to the CU size. In other words, CUs with different sizes will use different predictors. Similarly, a decoder can derive which predictor is used according to the PU size or other related unit size.

[0028] A predictor may contain the last palette size and a palette with NMax entries, where NMax can be equal to the size of palette and triplet palette.

[0029] For a palette coded (i.e., color index coded) CU, its palettes or triplet palette to be used can be derived from the predictor.

C. Palette and triplet palette sharing mode



[0030] Palette sharing mode can use palette predictors directly without modification. The palette to be used in the current CU is derived according to recorded last palette size. An example of palette and triplet palette sharing is shown in Fig. 2, where the last recorded palette size is 10. Accordingly, the first 10 entries in the palette predictor will be extracted directly to form a palette for the current CU.

[0031] Palette sharing mode can also signal a new size value. The palette is then derived from the palette predictor directly only with size modification.

D. Palette and triplet palette normal mode



[0032] In normal mode, information of palette modification will be signaled to indicate how to derive a palette from the palette predictor. Fig. 3 illustrates an example of palette predictor derivation. First, the new size is signaled to indicate the new palette size (e.g., 3) as shown in the "modification" table of Fig. 3. A bit to indicate will be signaled for each entry of the palette to indicate whether the value needs to be modified. If it needs to be modified, a new value will be further signaled. For example, New Value0 and New Value2 in Fig. 3 are used to modify major color indexes 0 and 2.

[0033] A prediction flag for each palette entry and each triplet palette entry can be signaled to indicate if the palette or triplet palette can reuse the previous coded palette or triplet palette for the corresponding entry, or the palette or triplet palette needs to be modified or transmitted. The prediction flag can be coded using context model based on context-adaptive binary arithmetic coding (CABAC) for the bin of the prediction flag.

[0034] The prediction flag can be limited to the first NP entries. In other words, for each entry smaller than NP, a prediction flag will be signaled. Otherwise, the entry will be forced to be modified. NP can be derived or transmitted.

[0035] To improve bitstream parsing throughout, all of the prediction flags or part of the prediction flags can be coded in the bypass mode. For example, the first N prediction flags are coded in the context coded mode and the rest prediction flags are coded in the bypass mode. N can be transmitted or derived.

[0036] Various examples of N for the first N palette entries are shown as follows for a system with the total number of entries in each color palette being 16 and the total number of entries for the triplet palette being 64:
  • N is 16 for the palette predictors and 64 for the triplet palette predictor, i.e. all flags are coded in the CABAC mode.
  • N is 0 for the palette predictors and triplet palette predictor, i.e. all flags are coded in the bypass mode.
  • N is 8 or 16 for the palette predictors.
  • N is 8, 16, 32, 48, or 64 for the triplet palette predictor.
  • N is derived based on the CU size.
  • N is derived based on the CU size. For example, when CU size is smaller than Sbypass, N is equal to N1. Otherwise, N is equal to N2.
    1. a. N2 is 16 for the palette predictors and 64 for the triplet palette predictor, i.e. all flags are coded in the CABAC mode. N1 is 0, 8, 16, 32, 48, or 64 for the palette predictors or the triplet palette predictor.
    2. b. Sbypass can be 8x8, 16x16, 32x32, or 64x64 CU.

Palette and Triplet Palette Predictors Updating



[0037] During decoding, the decoder maintains M sets of palettes predictors and triplet palette predictor. At the beginning of decoding each slice, the decoder can initialize predictors with transmitted or derived values.

[0038] After a palette coded CU is decoded, the predictors will be updated if the modification is signaled. If the palette size is not equal to the maximum size, the non-reused palette can be preserved in the table as palette predictors.

[0039] Before, during or after updating the palette predictor, the decoder can check for redundant index values in the predictor and the updated information. The decoder can remove redundant values if any redundant value is found. For example, after updating, the decoder can remove redundant values in the palette predictor. In another example, after updating and preserving the non-reused palette, the decoder can remove redundant values in the palette predictor. If the palette predictor size is larger than the maximum palette predictor size, the values not belonging to the maximum size of entries of the palette predictor can be dropped. The maximum size can be derived or transmitted. The derived position to be dropped can be the largest position.

[0040] Fig. 4 shows an example of predictor updating. During decoding the (N-1)th palette coded CU, a modification is signaled including New Size=3, modification flags and the new values of first and third entry. After decoding the (N-1)th palette coded CU, the predictor will be updated accordingly. The updated fields are indicated by patterned background in Fig. 4.

[0041] After a palette coded CU is decoded, the predictors can be updated if the modification is signaled.

[0042] The palette and/or triplet palette predictors can be reset at a slice-level, tile-level, LCU-row-level, or LCU-level. The reset can be turned on when Wavefront Parallel Processing (WPP) is applied. For the reset process, the palette and triplet palette predictors can be set to a predefined or derived value (e.g., 0) or can inherit the palette and triplet palette of an upper-right block of the current CU or PU.

[0043] An encoder can choose a triplet color to represent a pixel. The encoder first quantizes color values to color indexes on a component wise basis. The encoder can use color indexes (two or three) to find a match or similar color combination. The "match" color combination may correspond to all color indexes that are the same. The "match" color combination may also correspond to a color combination that some color indexes are the same and the different color indexes correspond to escape color indexes. In other word, the encoder can use an escape color to represent a major color. The "match" color combination may also correspond to the color indexes that are within the same quantization zone.

[0044] To calculate palettes and triplet palette, an encoder calculates the histogram of color values and chooses a set of representative major color values. These major colors are stored as component-wise palettes. Based on the major color values of the component-wise palettes, the encoder may further calculate the histogram of triplet colors and choose a set of representative color combinations (triplet colors). The combinations are recorded in a triplet (palette). If there are multiple matches or similar triplet colors, the encoder can choose a triplet color with most same major indexes.

[0045] To choose representative major colors or triplet colors, the encoder can sequentially choose the values with highest frequencies in the histogram within a quantization zone. When the histogram is reset, the histogram values within a quantization zone are reset to zeros, where "within a quantization zone" refers to the case that some color indexes are the same and those different color indexes in palette color correspond to escape color indexes. In other words, the encoder may use an escape color to represent a major color.

[0046] To choose representative major colors or triplet colors, the encoder can sequentially choose the major colors or triplet color having peak frequency in the histogram and reset the histogram values within the quantization zone to zero.

[0047] After the encoder decides major colors in the palette, the encoder can order colors in the palette according to the colors in the predictor. For example, color ordering for the palette can be similar to the color ordering for the predictor, and colors are placed in the same entry for both the palette and the predictor. This can minimize the modification required for palette predictor updating. The "similar" color refers to the same color or a color within a predefined or derived quantization step of the same color. If a color is "similar" to multiple colors in the palette predictor, the encoder can place the color at an entry with smallest index. Alternatively, the encoder can place the color at the entry with largest index.

[0048] In the case that a triplet color is used, after the encoder decides triplet colors in the triplet palette, the encoder can order the colors in the triplet palette according to the colors in the predictor. For example, color ordering for the triplet palette can be similar to the color ordering of the predictor, and a triplet is placed in the same entry for both the triplet palette and the predictor. The "similar" color refers to the case that some color indexes are the same and these colors with different color indexes in palette color correspond to escape color indexes. In other word, an encoder can use an escape color to represent a major color. The definition of "similar" can be that the color indexes are within the same quantization zone. If a color is similar to multiple colors in the predictor, the encoder can place the color at the entry with the smallest index. Alternatively, the encoder can order the color at the entry with the largest index.

[0049] The palette prediction and/or coding methods disclosed above can also be applied to other palette methods, such as these methods described in JCTVC-O0218, JCTVC-N0247 and JCTVC-O0182. A decoder can maintain palette predictors according to different palette structures.

[0050] For example, a decoder can maintain the palette predictors. Each entry of the predictor contains the color values of multiple color components. For example, the present invention can be applied to the method as disclosed in JCTVC-O0182. Accordingly, each entry contains 3 values for the RGB and YUV444 formats, and 2 values for UV in YUV422 and YUV420 formats, and 1 value for Y in YUV400, YUV422, and YUV420 formats.

[0051] Fig. 5A illustrates an example of current palette predictor derived from a last coded palette according an existing color index coding. The current palette contains four entries corresponding to Value0, Value3, Value2 and Value3 respectively. The current palette then becomes the next palette prediction according to the existing method. If the next palette which is a performance measurement commonly used for video coding.

[0052] Fig. 6 illustrates an exemplary flowchart of a system for color index coding using a palette predicted from a palette predictor including one or more entries form a previous palette predictor according to an embodiment of the present invention. The system receives input data associated with a current block as shown in step 610, where pixel values of the current block correspond to multiple color components. The input data corresponds to pixel data or color indices of the current block to be coded at an encoder side or corresponds to coded data or coded color indices of the current block to be decoded at the decoder side. The input data may be retrieved from memory (e.g., computer memory, buffer (RAM or DRAM) or other media) or from a processor. If a previous block is valid, first information related to a previous palette set corresponding to one or more previous palettes, one previous triplet-palette, or both associated with the previous block is received in step 620. Second information related to a previous palette predictor set corresponding to one or more previous palette predictors, one previous triplet-palette predictor, or both associated with the previous block is received in step 630. In one embodiment, there is at least one valid previous block associated with the current block as long as the current block is not at a first block in a slice, tile, largest coding unit row, or largest coding unit. A current palette predictor set corresponding to one or more current palette predictors, one current triplet-palette predictor, or both associated with the current block is derived based on the previous palette set and the previous palette predictor set by including at least one unused color or triplet of the previous palette predictor set in the current palette predictor set when at least one unused color or triplet of the previous palette predictor set in the current palette predictor set exists as shown in step 640. A current palette set corresponding to one or more current palettes, one current triplet-palette, or both associated with the current block is derived from the current palette predictor set as shown in step 650. Palette encoding or decoding is then applied to the current block according to the current palette set as shown in step 660.

[0053] The flowchart shown is intended to illustrate an example of color index coding using the palette predictor propagation across blocks according to the present invention. A person skilled in the art may modify each step, re-arranges the steps, split a step, or combine steps to practice the present invention without departing from the present invention. In the disclosure, specific syntax and semantics have been used to illustrate examples to implement embodiments of the present invention. A skilled person may practice the present invention by substituting the syntax and semantics with equivalent syntax and semantics without departing from the present invention.

[0054] The above description is presented to enable a person of ordinary skill in the art to practice the present invention as provided in the context of a particular application and its requirement. Various modifications to the described embodiments will be apparent to those with skill in the art, and the general principles defined herein may be applied to other embodiments. Therefore, the present invention is not intended to be limited to the particular embodiments shown and described, but is to be accorded the widest scope consistent with the principles and novel features herein disclosed. In the above detailed description, various specific details are illustrated in order to provide a thorough understanding of the present invention. Nevertheless, it will be understood by those skilled in the art that the present invention may be practiced.

[0055] Embodiment of the present invention as described above may be implemented in various hardware, software codes, or a combination of both. For example, an embodiment of the present invention can be one or more electronic circuits integrated into a video compression chip or program code integrated into video compression software to perform the processing described herein. An embodiment of the present invention may also be program code to be executed on a Digital Signal Processor (DSP) to perform the processing described herein. The invention may also involve a number of functions to be performed by a computer processor, a digital signal processor, a microprocessor, or field programmable gate array (FPGA). These processors can be configured to perform particular tasks according to the invention, by executing machine-readable software code or firmware code that defines the particular methods embodied by the invention. The software code or firmware code may be developed in different programming languages and different formats or styles. The software code may also be compiled for different target platforms. However, different code formats, styles and languages of software codes and other means of configuring code to perform the tasks in accordance with the invention will not depart from the scope of the invention.

[0056] The invention may be embodied in other specific forms without departing from its essential characteristics. The described examples are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is therefore, indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.


Claims

1. A method of coding a block of video data using palette coding for a video coding system, the method comprising:

receiving input data (610) associated with a first block corresponding to a current block, wherein pixel values of the first block correspond to multiple color components;

if the first block is different from the beginning block in a slice, largest coding unit row, or largest coding unit:

receiving first information (620) related to a second palette set corresponding to one or more second palettes, one second triplet-palette, or both associated with a second block which is coded immediately prior to the first block;

receiving second information (630) related to a second palette predictor set corresponding to one or more second palette predictors, one second triplet-palette predictor, or both associated with the second block;

deriving a first palette predictor set (640) corresponding to one or more first palette predictors, one first triplet-palette predictor, or both associated with the first block based on the second palette set and the second palette predictor set by including at least one non-reused entry of the second palette predictor set into the first palette predictor set, when said at least one non-reused entry exists, wherein said non-reused entry corresponds to one color or triplet of the second palette predictor set, that is unused in the second palette set;

deriving a first palette set (650) corresponding to one or more first palettes, one first triplet-palette, or both associated with the first block from the first palette predictor set; and

applying palette encoding or decoding to the first block (660) according to the first palette set;

characterized in that

if a palette size of said second palette set is smaller than a maximum palette size, then a plurality of last non-reused entries of said second palette predictor set are copied to corresponding entries of said first palette predictor set according to a palette-index order of the second palette predictor set, wherein only the plurality of last non-reused entries of said second palette predictor set having a palette index which is higher than or equal to a palette size of said second palette predictor set are copied.


 
2. The method of Claim 1, further comprising signaling one or more first updating signals for updating the first palette predictor set based on the second palette set and the second palette predictor set.
 
3. The method of Claim 2, wherein if a palette predictor size of said one or more first palette predictors, said one first triplet-palette predictor, or both is larger than a maximum palette predictor size after updating, then the predictor entries of said one or more first palette predictors, said one first triplet-palette predictor, or both, which exceed said maximum palette predictor size, are dropped, and wherein said maximum palette predictor size is derived or transmitted.
 
4. The method of any of Claims 1 to 3, wherein a number of second palette predictor sets, which are maintained at a decoder side, corresponds to M, wherein each of the maintained second palette predictor sets corresponds to one or more second palettes, one second triplet-palette, or both, and M is an integer greater than zero.
 
5. The method of any of the Claims 1 to 4, wherein the second block and the first block correspond to coding units or prediction units, preferably,
wherein said deriving the first palette predictor set (640) depends on coding-unit size if the second block and the first block correspond to the coding units, or said deriving the first palette predictor set depends on prediction-unit size if the second block and the first block correspond to the prediction units.
 
6. The method of any of the Claims 1 to 5, wherein said deriving the first palette predictor set (640) comprises updating the second palette predictor set according to one or more updating signals, preferably:

wherein said one or more updating signals include a last palette-size or a new palette-size, wherein the first palette set is derived from the first palette predictor set by copying a number of beginning entries of the first palette predictor set, and the number corresponds to the last palette-size or the new palette-size, or

wherein said one or more updating signals include new palette-size for each of said one or more first palettes, said one first triplet-palette, or both associated with the first block, a number of modification flags corresponding to new palette-size to identify whether a number of entries of each of said one or more first palettes, said one first triplet-palette, or both associated with the first block are updated, and a new value for each to-be-modified entry as indicated by a corresponding modification flag, and the new palette-size is an integer, and, preferably,

wherein at least one modification flag is context adaptive coded using context-adaptive binary arithmetic coding (CABAC) or
context adaptive coded using a bypass mode of context-adaptive binary arithmetic coding (CABAC).


 
7. The method of Claim 6, wherein modification flags are used for beginning NP entries of each of said one or more first palettes, said one first triplet-palette, or both associated with the first block, and remaining entries are forced to be modified, where NP is a positive integer.
 
8. The method of any of the Claims 1 to 7, wherein said deriving the first palette predictor set comprises checking and removing redundancy in the first palette predictor set, preferably, wherein said checking and removing redundancy is performed before, during or after one or more new entries are placed into the first palette predictor set.
 
9. The method of Claim 1, wherein if the first block corresponds to the beginning block in a slice, largest coding unit row, or largest coding unit, the current palette set is derived by an encoder without the second palette set nor the second palette predictor set.
 
10. The method of Claim 9, wherein the encoder quantizes color values of the first block to color indexes for each color component and assigns two or three color indexes associated with one pixel of the first block to one triplet index representing a two-color index or a three-color index; wherein said two or three color indexes associated with said one pixel of the first block are mapped to said one triplet index according to the first palette set if said two or three color indexes associated with said one pixel of the first block are the same as a color combination of said one triplet index,
or if one part of said two or three color indexes associated with said one pixel of the first block is the same as a corresponding part of the color combination of said one triplet index and a remaining part of said two or three color indexes associated with said one pixel of the first block corresponds to one or more escape color indexes.
 
11. The method of Claim 10, wherein the encoder determines a first histogram of the color values and selects a first set of representative colors as major colors for each color component, wherein the major colors for each color component are stored as the first palette associated with the respective color component; and the encoder determines, based on the major colors of the first palettes associated with the color components, a second histogram of triplet colors and selects a second set of representative color combinations as triplet colors which are recorded in the first triplet palette.
 
12. An apparatus for coding a block of video data using palette coding for a video coding system, the apparatus comprising one or more electronic circuits configured to implement all the steps of a method according to any of the previous claims.
 


Ansprüche

1. Verfahren eines Codierens eines Blocks von Videodaten unter Verwendung einer Palettencodierung für ein Videocodierungssystem, wobei das Verfahren umfasst:

Empfangen von Eingangsdaten (610), die mit einem ersten Block verknüpft sind, der zu einem aktuelle Block korrespondiert, wobei Pixelwerte des ersten Blocks zu mehreren Farbkomponenten korrespondieren;

wenn der erste Block von dem Anfangsblock in einer Slice, einer größten Codierungseinheitenzeile oder einer größten Codierungseinheit verschieden ist:

Empfangen erster Informationen (620), die sich auf einen zweiten Palettensatz beziehen, der zu einer oder mehreren zweiten Paletten, einer zweiten Dreiergruppenpalette oder beiden korrespondiert, die mit einem zweiten Block verknüpft sind, welcher unmittelbar vor dem ersten Block codiert wird;

Empfangen zweiter Informationen (630), die sich auf einen zweiten Palettenprädiktorsatz beziehen, der zu einem oder mehreren zweiten Palettenprädiktoren, einem zweiten Dreiergruppenpalettenprädiktor oder beiden korrespondiert, die mit dem zweiten Block verknüpft sind;

Ableiten eines ersten Palettenprädiktorsatzes (640), der zu einem oder mehreren ersten Palettenprädiktoren, einem ersten Dreiergruppenpalettenprädiktor oder beiden korrespondiert, die mit dem ersten Block verknüpft sind, basierend auf dem zweiten Palettensatz und dem zweiten Palettenprädiktorsatz durch Einschließen mindestens eines nicht wiederverwendeten Eintrags des zweiten Palettenprädiktorsatzes in den ersten Palettenprädiktorsatz, wenn der mindestens eine nicht wiederverwendete Eintrag vorhanden ist, wobei der nicht wiederverwendete Eintrag zu einer Farbe oder einer Dreiergruppe des zweiten Palettenprädiktorsatzes korrespondiert, welcher in dem zweiten Palettensatz nicht verwendet wird;

Ableiten eines ersten Palettensatzes (650), der zu einer oder mehreren ersten Paletten, einer ersten Dreiergruppenpalette oder beiden korrespondiert, die mit dem ersten Block verknüpft sind, von dem ersten Palettenprädiktorsatz; und

Anwenden einer Palettencodierung oder -decodierung auf den ersten Block (660) gemäß dem ersten Palettensatz;

dadurch gekennzeichnet, dass

wenn eine Palettengröße des zweiten Palettensatzes kleiner ist als eine maximale Palettengröße, dann wird eine Mehrzahl von letzten nicht wiederverwendeten Einträgen des zweiten Palettenprädiktorsatzes zu korrespondierenden Einträgen des ersten Palettenprädiktorsatzes gemäß einer Palettenindexfolge des zweiten Palettenprädiktorsatzes kopiert, wobei nur die Mehrzahl von letzten nicht wiederverwendeten Einträgen des zweiten Palettenprädiktorsatzes kopiert wird, die einen Palettenindex aufweisen, welcher höher oder gleich einer Palettengröße des zweiten Palettenprädiktorsatzes ist.


 
2. Verfahren gemäß Anspruch 1, weiter umfassend ein Signalisieren eines oder mehrerer Aktualisierungssignale zum Aktualisieren des ersten Palettenprädiktorsatzes basierend auf dem zweiten Palettensatz und dem zweiten Palettenprädiktorsatz.
 
3. Verfahren gemäß Anspruch 2, wobei, wenn eine Palettenprädiktorgröße des einen oder der mehreren ersten Palettenprädiktoren, des einen ersten Dreiergruppenpalettenprädiktors oder beider größer ist als eine maximale Palettenprädiktorgröße nach einer Aktualisierung, dann werden die Prädiktoreinträge des einen oder der mehreren ersten Palettenprädiktoren, des einen ersten Dreiergruppenpalettenprädiktors oder beider, welche die maximale Palettenprädiktorgröße überschreiten, fallengelassen, und wobei die maximale Palettenprädiktorgröße abgeleitet oder übertragen wird.
 
4. Verfahren gemäß einem der Ansprüche 1 bis 3, wobei eine Anzahl von zweiten Palettenprädiktorsätzen, welche auf einer Decodiererseite gepflegt werden, zu M korrespondiert, wobei jeder der gepflegten zweiten Palettenprädiktorsätze zu einer oder mehreren zweiten Paletten, einer zweiten Dreiergruppenpalette oder beiden korrespondiert, und M eine ganze Zahl größer als null ist.
 
5. Verfahren gemäß einem der Ansprüche 1 bis 4, wobei der zweite Block und der erste Block zu Codierungseinheiten oder Prädiktionseinheiten korrespondieren,
wobei bevorzugt das Ableiten des ersten Palettenprädiktorsatzes (640) von einer Codierungseinheitengröße abhängt, wenn der zweite Block und der erste Block zu den Codierungseinheiten korrespondieren, oder das Ableiten des ersten Palettenprädiktorsatzes von einer Prädiktionseinheitengröße abhängt, wenn der zweite Block und der erste Block zu den Prädiktionseinheiten korrespondieren.
 
6. Verfahren gemäß einem der Ansprüche 1 bis 5, wobei das Ableiten des ersten Palettenprädiktorsatzes (640) ein Aktualisieren des zweiten Palettenprädiktorsatzes gemäß einem oder mehreren Aktualisierungssignalen umfasst, wobei bevorzugt:

das eine oder die mehreren Aktualisierungssignale eine letzte Palettengröße oder eine neue Palettengröße aufweisen, wobei der erste Palettensatz von dem ersten Palettenprädiktorsatz durch ein Kopieren einer Anzahl von Anfangseinträgen des ersten Palettenprädiktorsatzes abgeleitet wird, und die Anzahl zu der letzten Palettengröße oder der neuen Palettengröße korrespondiert, oder

das eine oder die mehreren Aktualisierungssignale eine neue Palettengröße für jede der einen oder mehreren ersten Paletten, der einen ersten Dreiergruppenpalette oder beide aufweisen, die mit dem ersten Block verknüpft sind, eine Anzahl von Modifikationskennzeichen, die zu einer neuen Palettengröße korrespondieren, um zu identifizieren, ob eine Anzahl von Einträgen jeder der einen oder mehreren ersten Paletten, der einen ersten Dreiergruppenpalette oder beider, die mit dem ersten Block verknüpft sind, aktualisiert werden, und einen neuen Wert für jeden zu modifizierenden Eintrag, wie durch ein korrespondierendes Modifikationskennzeichen angezeigt, aufweisen, und die neue Palettengröße eine ganze Zahl ist, und bevorzugt

wobei mindestens ein Modifikationskennzeichen kontext-adaptiv unter Verwendung einer kontext-adaptiven binären arithmetischen Codierung (CABAC) codiert wird oder kontext-adaptiv unter Verwendung eines Bypass-Modus einer kontext-adaptiven binären arithmetischen Codierung (CABAC) codiert wird.


 
7. Verfahren gemäß Anspruch 6, wobei Modifikationskennzeichen für NP Anfangseinträge jeder der einen oder mehreren ersten Paletten, der einen ersten Dreiergruppenpalette oder beider, die mit dem ersten Block verknüpft sind, verwendet werden, und übrige Einträge erzwungen modifiziert werden, wobei NP eine positive ganze Zahl ist.
 
8. Verfahren gemäß einem der Ansprüche 1 bis 7, wobei das Ableiten des ersten Palettenprädiktorsatzes ein Prüfen und Entfernen von Redundanz in dem ersten Palettenprädiktorsatz umfasst, wobei bevorzugt das Prüfen und Entfernen von Redundanz ausgeführt wird, bevor, während oder nachdem ein oder mehrere neue Einträge in den ersten Palettenprädiktorsatz eingebracht werden.
 
9. Verfahren gemäß Anspruch 1, wobei, wenn der ersten Block zu dem Anfangsblock in einer Slice, einer größten Codierungseinheitenzeile oder einer größten Codierungseinheit korrespondiert, der aktuelle Palettensatz durch einen Codierer ohne den zweiten Palettensatz und auch ohne den zweiten Palettenprädiktorsatz festgelegt wird.
 
10. Verfahren gemäß Anspruch 9, wobei der Codierer Farbwerte des ersten Blocks zu Farbindizes für jede Farbkomponente quantisiert und zwei oder drei Farbindizes, die mit einem Pixel des ersten Blocks verknüpft sind, zu einem Dreiergruppenindex zuweist, der einen Zweifarbindex oder einen Dreifarbindex repräsentiert; wobei die zwei oder drei Farbindizes, die mit dem einen Pixel des ersten Blocks verknüpft sind, gemäß dem ersten Palettensatz auf den einen Dreiergruppenindex abgebildet werden, wenn die zwei oder drei Farbindizes, die mit dem einen Pixel des ersten Blocks verknüpft sind, die gleichen sind wie eine Farbkombination des einen Dreiergruppenindex,
oder wenn ein Teil der zwei oder drei Farbindizes, die mit dem einen Pixel des ersten Blocks verknüpft sind, der gleiche ist wie ein korrespondierender Teil der Farbkombination des einen Dreiergruppenindex, und ein übriger Teil der zwei oder drei Farbindizes, die mit dem einen Pixel des ersten Blocks verknüpft sind, zu einem oder mehreren Ausweichfarbindizes korrespondiert.
 
11. Verfahren gemäß Anspruch 10, wobei der Codierer ein erstes Histogramm der Farbwerte bestimmt und einen ersten Satz von repräsentativen Farben als Hauptfarben für jede Farbkomponente auswählt, wobei die Hauptfarben für jede Farbkomponente als die erste Farbpalette gespeichert werden, die mit der jeweiligen Farbkomponente verknüpft ist; und der Codierer basierend auf den Hauptfarben der ersten Paletten, die mit den Farbkomponenten verknüpft sind, ein zweites Histogramm von Dreiergruppenfarben bestimmt und einen zweiten Satz von repräsentativen Farbkombinationen als Dreiergruppenfarben auswählt, welche in der ersten Dreiergruppenpalette aufgezeichnet werden.
 
12. Vorrichtung zum Codieren eines Blocks von Videodaten unter Verwendung einer Palettencodierung für ein Videocodierungssystem, wobei die Vorrichtung eine oder mehrere elektronische Schaltungen aufweist, die ausgelegt sind, alle Schritte eines Verfahrens gemäß einem der vorstehenden Ansprüche zu implementieren.
 


Revendications

1. Procédé de codage d'un bloc de données vidéo en utilisant un codage de palette pour un système de codage vidéo, le procédé comprenant le fait :

de recevoir des données d'entrée (610) associées à un premier bloc correspondant à un bloc actuel, où des valeurs de pixels du premier bloc correspondent à de multiples composantes de couleur ;

si le premier bloc est différent du bloc de démarrage dans une tranche, la plus grande rangée d'unités de codage ou la plus grande unité de codage :

de recevoir des premières informations (620) relatives à un deuxième ensemble de palettes correspondant à au moins une deuxième palette, à une deuxième palette de triplets, ou aux deux associées à un deuxième bloc qui est codé immédiatement avant le premier bloc ;

de recevoir des deuxièmes informations (630) relatives à un deuxième ensemble de prédicteurs de palette correspondant à au moins un deuxième prédicteur de palette, à un deuxième prédicteur de palette de triplets, ou aux deux associés au deuxième bloc ;

d'obtenir un premier ensemble de prédicteurs de palette (640) correspondant à au moins un premier prédicteur de palette, à un premier prédicteur de palette de triplets, ou aux deux associés au premier bloc sur la base du deuxième ensemble de palettes et du deuxième ensemble de prédicteurs de palette en incluant au moins une entrée à usage unique du deuxième ensemble de prédicteurs de palette dans le premier ensemble de prédicteurs de palette, lorsque ladite au moins une entrée à usage unique existe, où ladite entrée à usage unique correspond à une couleur ou un triplet du deuxième ensemble de prédicteurs de palette, qui n'est pas utilisé dans le deuxième ensemble de palettes ;

d'obtenir un premier ensemble de palettes (650) correspondant à au moins une première palette, à une première palette de triplets, ou aux deux associées au premier bloc à partir du premier ensemble de prédicteurs de palette ; et

d'appliquer un codage ou un décodage de palette au premier bloc (660) conformément au premier ensemble de palettes ;

caractérisé en ce que

si une taille de palette dudit deuxième ensemble de palettes est inférieure à une taille de palette maximale, alors une pluralité de dernières entrées à usage unique dudit deuxième ensemble de prédicteurs de palette sont copiées dans des entrées correspondantes dudit premier ensemble de prédicteurs de palette conformément à un ordre d'indice de palette du deuxième ensemble de prédicteurs de palette, où seule la pluralité de dernières entrées à usage unique dudit deuxième ensemble de prédicteurs de palette ayant un indice de palette qui est supérieur ou égal à une taille de palette dudit deuxième ensemble de prédicteurs de palette sont copiées.


 
2. Procédé de la revendication 1, comprenant en outre la signalisation d'au moins un premier signal de mise à jour pour mettre à jour le premier ensemble de prédicteurs de palette sur la base du deuxième ensemble de palettes et du deuxième ensemble de prédicteurs de palette.
 
3. Procédé de la revendication 2, dans lequel si une taille de prédicteur de palette dudit au moins un premier prédicteur de palette, dudit premier prédicteur de palette de triplets, ou des deux, est supérieure à une taille de prédicteur de palette maximale après la mise à jour, alors les entrées de prédicteur dudit premier prédicteur de palette, dudit premier prédicteur de palette de triplets, ou des deux, qui dépassent ladite taille de prédicteur de palette maximale, sont abandonnées, et où ladite taille de prédicteur de palette maximale est obtenue ou transmise.
 
4. Procédé de l'une des revendications 1 à 3, dans lequel un certain nombre de deuxièmes ensembles de prédicteurs de palette, qui sont maintenus au niveau d'un côté décodeur, correspond à M, où chacun des deuxièmes ensembles de prédicteurs de palette maintenus correspond à au moins une deuxième palette, à une deuxième palette de triplets, ou aux deux, et M est un nombre entier supérieur à zéro.
 
5. Procédé de l'une des revendications 1 à 4, dans lequel le deuxième bloc et le premier bloc correspondent à des unités de codage ou des unités de prédiction, de préférence,
dans lequel ladite obtention du premier ensemble de prédicteurs de palette (640) dépend de la taille d'unité de codage si le deuxième bloc et le premier bloc correspondent aux unités de codage, ou ladite obtention du premier ensemble de prédicteurs de palette dépend de la taille d'unité de prédiction si le deuxième bloc et le premier bloc correspondent aux unités de prédiction.
 
6. Procédé de l'une des revendications 1 à 5, dans lequel ladite obtention du premier ensemble de prédicteurs de palette (640) comprend la mise à jour du deuxième ensemble de prédicteurs de palette conformément à au moins un signal de mise à jour, de préférence :

dans lequel ledit au moins un signal de mise à jour comporte une dernière taille de palette ou une nouvelle taille de palette, où le premier ensemble de palettes est obtenu à partir du premier ensemble de prédicteurs de palette en copiant un certain nombre d'entrées de démarrage du premier ensemble de prédicteurs de palette, et le nombre correspond à la dernière taille de palette ou à la nouvelle taille de palette, ou

dans lequel ledit au moins un signal de mise à jour comporte une nouvelle taille de palette pour chacune de ladite au moins une première palette, de ladite première palette de triplets, ou des deux associées au premier bloc, un certain nombre de drapeaux de modification correspondant à la nouvelle taille de palette pour identifier si un certain nombre d'entrées de chacune de ladite au moins une première palette, de ladite première palette de triplets, ou des deux associées au premier bloc sont mises à jour, et une nouvelle valeur pour chaque entrée à modifier comme l'indique un drapeau de modification correspondant, et la nouvelle taille de palette est un nombre entier, et, de préférence,

dans lequel au moins un drapeau de modification est codé de manière adaptative au contexte en utilisant un codage arithmétique binaire adaptatif au contexte (CABAC) ou codé de manière adaptative au contexte en utilisant un mode de contournement de codage arithmétique binaire adaptatif au contexte (CABAC).


 
7. Procédé de la revendication 6, dans lequel des drapeaux de modification sont utilisés pour des entrées de démarrage NP de chacune de ladite au moins une première palette, de ladite première palette de triplets, ou des deux associées au premier bloc, et les entrées restantes sont contraintes à être modifiées, où NP est un nombre entier positif.
 
8. Procédé de l'une des revendications 1 à 7, dans lequel ladite obtention du premier ensemble de prédicteurs de palette comprend la vérification et la suppression de redondance dans le premier ensemble de prédicteurs de palette, de préférence, où ladite vérification et ladite suppression de redondance sont effectuées avant, pendant ou après la mise en place d'au moins une nouvelle entrée dans le premier ensemble de prédicteurs de palette.
 
9. Procédé de la revendication 1, dans lequel si le premier bloc correspond au bloc de démarrage dans une tranche, la plus grande rangée d'unités de codage ou la plus grande unité de codage, l'ensemble de palettes actuel est obtenu par un codeur sans le deuxième ensemble de palettes ni le deuxième ensemble de prédicteurs de palette.
 
10. Procédé de la revendication 9, dans lequel le codeur quantifie des valeurs de couleur du premier bloc selon des indices de couleur pour chaque composante de couleur et attribue deux ou trois indices de couleur associés à un pixel du premier bloc à un indice de triplet représentant un indice de deux couleurs ou un indice de trois couleurs ; dans lequel lesdits deux ou trois indices de couleur associés audit pixel du premier bloc sont mis en correspondance avec ledit indice de triplet conformément au premier ensemble de palettes si lesdits deux ou trois indices de couleur associés audit pixel du premier bloc sont identiques à une combinaison de couleurs dudit indice de triplet,
ou si une partie desdits deux ou trois indices de couleur associés audit pixel du premier bloc est identique à une partie correspondante de la combinaison de couleurs dudit indice de triplet et une partie restante desdits deux ou trois indices de couleur associés audit pixel du premier bloc correspond à au moins un indice de couleur d'échappement.
 
11. Procédé de la revendication 10, dans lequel le codeur détermine un premier histogramme des valeurs de couleur et sélectionne un premier ensemble de couleurs représentatives en tant que couleurs principales pour chaque composante de couleur, où les couleurs principales pour chaque composante de couleur sont stockées comme étant la première palette associée à la composante de couleur respective ; et le codeur détermine, sur la base des couleurs principales des premières palettes associées aux composantes de couleur, un deuxième histogramme de couleurs de triplets et sélectionne un deuxième ensemble de combinaisons de couleurs représentatives en tant que couleurs de triplets qui sont enregistrées dans la première palette de triplets.
 
12. Appareil pour coder un bloc de données vidéo en utilisant un codage de palette pour un système de codage vidéo, l'appareil comprenant au moins un circuit électronique configuré pour mettre en Ĺ“uvre toutes les étapes d'un procédé selon l'une des revendications précédentes.
 




Drawing


















REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description