Palette Predictor Set Construction for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing color index coding methods for video data, such as those in the HEVC standard, face inefficiencies in palette prediction and sharing across coding units, leading to suboptimal bitrate reduction and coding performance, especially for screen content with limited color variations.
Innovation Solution
A method that derives a current palette predictor set by incorporating unused colors or triplets from a previous palette predictor set, allowing for adaptive palette encoding and decoding across blocks, with mechanisms for updating and redundancy removal, and employing layered color representation to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If palette prediction is performed using only previously transmitted palette entries, then the palette construction is simple, but the coding efficiency is suboptimal due to inability to utilize unused palette capacity
Solution Approach 1:
The patent changes the parameter of palette entry selection from only previously transmitted entries to include both previously transmitted entries and unused palette entries. This allows the current palette predictor set to be constructed by selecting entries from the previous palette predictor set that were not used in the current block, thereby improving coding efficiency by utilizing available palette capacity while maintaining manageable complexity through systematic selection rules
Solution Approach 2:
The patent introduces dynamic palette construction where the palette predictor set is adaptively updated based on usage patterns. The encoder and decoder dynamically identify which palette entries were used and which remain unused, then selectively carry forward unused entries to the next palette predictor set. This dynamic adaptation improves coding efficiency without requiring complex manual intervention
2Adaptability or versatility
If the palette size is increased to accommodate more colors, then the color representation capability improves, but the bitrate increases due to more data to transmit
Solution Approach 1:
The patent applies partial action by transmitting only the necessary subset of palette entries rather than the entire palette. The encoder identifies and transmits only those palette entries that are actually used in the current block, leaving unused entries implicit. This reduces the quantity of transmitted data (bitrate) while maintaining full color representation capability for the used entries
Solution Approach 2:
The patent extracts and transmits only the relevant portion of the palette data - specifically, only the entries that are actually used in the current block. By separating used entries from unused entries and transmitting only the used ones, the patent reduces bitrate while preserving complete color representation capability for the necessary colors
3Quantity of substance
If palette entries are reused across blocks without updating, then the bitrate is reduced, but the coding accuracy deteriorates due to redundancy
Solution Approach 1:
The patent performs preliminary identification of unused palette entries before constructing the current palette predictor set. By proactively identifying which entries from the previous palette predictor set were not used in the current block, the system prepares an optimized palette predictor set in advance that eliminates redundancy. This preliminary action maintains low bitrate while improving coding accuracy by preventing redundant transmissions
4Measurement precision
If the system transmits complete palette information for each block, then the decoding accuracy is high, but the bitrate efficiency is poor
Solution Approach 1:
The patent uses copying by having the decoder reconstruct unused palette entries from the previous palette predictor set without receiving them in the bitstream. Instead of transmitting complete palette information, the system copies the necessary unused entries from the previously stored palette predictor set at the decoder side, maintaining decoding accuracy while significantly reducing bitrate
Data Source
Figure 1A~1B
Figure 2~3
Figure 4~5A
AI summary
A method of coding a block of video data using palette coding is disclosed. According to the present invention, 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. A current palette set corresponding to current palettes/triplet-palette associated with the current block is derived from the current palette predictor set for palette encoding or decoding.