Palette Predictor Update for Screen Content Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current palette coding techniques in HEVC are limited to coding units and do not efficiently utilize sub-blocks or prediction units, leading to suboptimal coding efficiency for screen content materials.

Innovation Solution

The method involves applying palette coding to sub-blocks or prediction units within a coding unit, allowing each sub-block to use its own palette table and enabling more flexible partition modes, such as NxN, 2NxM, and asymmetric partitions, while skipping quantization matrices when residuals are not transformed into the frequency domain.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If palette coding is applied to the entire coding unit, then coding efficiency is improved, but prediction accuracy is insufficient due to lack of local adaptation

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprediction accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides the coding unit into multiple sub-blocks, with each sub-block having its own palette table and prediction parameters. This segmentation allows local adaptation to different color distributions within the coding unit, improving prediction accuracy while maintaining overall coding efficiency through parallel processing of sub-blocks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different palette tables and prediction modes to different sub-blocks based on their local color characteristics. Each sub-block can use its own optimized palette, allowing the coding system to adapt to local variations in screen content while maintaining high coding efficiency across the entire coding unit.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If each sub-block uses its own palette table, then prediction accuracy is improved, but device complexity increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidpalette management complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges palette tables from multiple sub-blocks to form a combined palette for the entire coding unit. This reduces the total number of unique colors that need to be encoded in the bitstream, simplifying palette management while still allowing each sub-block to use its own optimized palette for prediction, thus balancing prediction accuracy with reduced complexity.

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If quantization matrices are always applied, then coding completeness is maintained, but bitstream size increases unnecessarily for non-transformed residuals

Engineering Contradiction:
Improvecoding completenessVSAvoidbitstream size
Core Design Contradiction:
ReliabilityVSLoss of substance

Solution Approach 1:

The patent applies quantization matrices only when residual transformation is performed, and skips quantization when residuals are not transformed. This partial application of quantization avoids unnecessary processing steps and reduces bitstream size while maintaining coding completeness for cases where transformation is beneficial, effectively removing excessive actions that do not contribute to coding performance.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3198875B1Method for palette predictor update
Publication Date: 2021.04.21 MEDIATEK INC
  • EP3198875B1 patent drawingFigure 1
  • EP3198875B1 patent drawingFigure 2A~2C
  • EP3198875B1 patent drawingFigure 3

AI summary

A method of palette coding to apply the palette coding to sub-blocks of a coding unit and to allow each sub-block to use an individual palette table is disclosed. If the current coding block is not partitioned, the palette coding is applied to the current coding block using a first palette. If the current coding block is partitioned into multiple sub-blocks, the palette coding is applied to each sub-block using an individual second palette. Each sub-block may correspond to one prediction unit (PU). In another embodiment of the present invention, a method is disclosed that skips update of palette predictor such as last coded palette table, last coded palette size and palette predictor size associated with the current coding block if the current palette size is smaller than or equal to palette update size.