Palette Predictor Encoding Reducing Signalling Costs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The Palette mode in HEVC encoding requires transmitting a large amount of data for each coding unit, leading to high signalling costs due to the need to transmit entire palettes for each block of pixels, which is inefficient and costly.
Innovation Solution
The method predicts the current palette from entries of two or more previously used palettes, using a bitmap of flags to select entries from one palette to enrich the current palette, reducing the amount of information needed to be transmitted by leveraging redundancy between consecutive blocks of pixels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If entire palettes are transmitted for each block of pixels, then prediction accuracy is maintained, but signalling costs become excessively high
Solution Approach 1:
The patent applies preliminary action by pre-establishing a palette predictor from previously processed blocks before encoding the current block. The palette predictor is built in advance using entries from previous palettes, so that when encoding the current block, only the difference (residue) needs to be transmitted rather than the entire palette. This reduces signalling costs while maintaining prediction accuracy through the pre-computed predictor.
Solution Approach 2:
The patent uses copying by replicating palette entries from previous blocks to form the palette predictor for the current block. Instead of transmitting all palette entries, the system copies relevant entries from previously processed blocks that have similar color characteristics. This copying approach reduces the amount of data to be transmitted while preserving the essential color information needed for accurate prediction.
2Loss of information
If palettes are predicted from previous palettes, then signalling costs are reduced, but the complexity of the encoding process increases
Solution Approach 1:
The patent applies segmentation by dividing the palette into individual entries and processing them separately. Instead of handling the entire palette as a single unit, the system segments the palette into discrete color entries that can be individually copied from previous blocks. This segmentation allows for more flexible and efficient palette prediction, reducing signalling costs while managing complexity through structured processing of individual palette components.
Solution Approach 2:
The patent uses parameter changes by modifying the representation of palette data from complete transmitted palettes to predicted palettes with only differences encoded. The system changes the fundamental parameter of how palette information is handled - from transmitting full palettes to transmitting only the residue or difference from the predicted palette. This parameter change significantly reduces signalling costs while the complexity is managed through algorithmic prediction rather than data transmission.
3Loss of information
If palettes are predicted using bitmap flags to select entries, then data transmission is optimized, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively copying only the necessary palette entries from previous blocks rather than copying all entries. The system uses bitmap flags to indicate which palette entries need to be copied, allowing for partial copying of only the relevant color information. This partial action approach optimizes data transmission by sending only essential information while the processing time is managed through efficient selective copying rather than full palette processing.
Data Source
AI summary
A method and a device for processing a current pixel block of an image using a palette prediction mode according to HEVC Format Range Extension. The mode uses a current palette to build a predictor block of indexes to predict the current pixel block. The current palette comprises entries associating entry indexes with pixel values. The method comprises predicting the current palette from entries of two or more palettes that are palettes previously used to process blocks of pixels. A palette predictor may be used that is built from the two or more palettes, preferably from the last used palette for a coding unit in the current coding entity and from the previously used palette for which a flag bitmap indicates whether or not its elements have been copied into the last used palette. Accordingly, the coding of the palette mode is improved.