Adaptive Search Range for DSC Block Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video compression techniques, such as those in the DSC standard, struggle to effectively compress highly complex regions in video data, as existing intra prediction methods are limited in providing satisfactory predictions for blocks similar to previous coded blocks, leading to inefficiencies in bandwidth usage and computational complexity.

Innovation Solution

The implementation of adaptive search range selection for coding blocks in block prediction mode, allowing the encoder to determine optimal search ranges based on rate distortion analysis and the location of blocks, which reduces decoder complexity and improves coding efficiency by allowing multiple and adaptive search ranges for superior candidate partition selection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a fixed search range is used for block prediction, then decoder complexity is reduced, but coding efficiency deteriorates due to inability to find superior candidate partitions

Engineering Contradiction:
Improvedecoder complexityVSAvoidcoding efficiency
Core Design Contradiction:
Device complexityVSProductivity

Solution Approach 1:

The patent applies dynamics by making the search range adaptive rather than fixed. The encoder determines optimal search ranges dynamically based on rate distortion analysis and block location, allowing the search range to vary across different blocks and coding scenarios. This resolves the contradiction by enabling the decoder to maintain low complexity (using a predetermined range) while the encoder achieves high coding efficiency through adaptive range selection.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of search range from a fixed value to an adaptive parameter determined by rate distortion analysis. By analyzing the relationship between compression rate and distortion for different search ranges, the system selects optimal ranges that maximize coding efficiency while keeping decoder complexity manageable through predetermined range structures.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If multiple adaptive search ranges are used for block prediction, then coding efficiency is improved by finding superior candidate partitions, but decoder complexity increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoiddecoder complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the search space into multiple predetermined ranges with different characteristics. Each range is optimized for specific block locations or coding scenarios, allowing the encoder to select the most appropriate range while the decoder can efficiently search within predetermined boundaries. This segmentation maintains decoder simplicity while enabling encoder-level adaptivity for improved coding efficiency.

Inventive Principle:
Principle #1Segmentation

3Ease of manufacture

If existing intra prediction methods are used, then implementation is simple and inexpensive, but compression performance deteriorates for highly complex video regions

Engineering Contradiction:
Improveimplementation simplicityVSAvoidcompression performance
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent performs preliminary rate distortion analysis to determine optimal search ranges before actual block prediction. By pre-analyzing the characteristics of complex video regions and determining appropriate search strategies in advance, the system prepares optimization parameters that can be applied during encoding. This preliminary action maintains implementation simplicity while significantly improving compression performance for complex regions.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3387832B1Multi-region search range for block prediction mode for display stream compression (DSC)
Publication Date: 2020.04.22 QUALCOMM INC
  • EP3387832B1 patent drawingFigure 1A
  • EP3387832B1 patent drawingFigure 1B
  • EP3387832B1 patent drawingFigure 2A

AI summary

A method for coding a block of video data in simplified block prediction mode of a constant bitrate video coding scheme for transmission over display links is disclosed. In one aspect, the method includes determining a candidate block to be used to predict a current block in a current slice, the candidate block being within a range of pixel positions that each correspond to a reconstructed pixel in the current slice. The range of pixel positions may comprise (i) a first region including one or more first pixel positions in a first line of pixels that overlaps the current block, and (ii) a second region including one or more second pixel positions in a second line of pixels that does not overlap the current block. The method may further comprise determining and signaling a prediction vector indicative of a pixel position of the candidate block.