Spatial Prediction Mode Redundancy Elimination in Image Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing spatial prediction methods in image coding, such as those in the H.264/AVC standard, suffer from inefficiencies due to redundant prediction modes, which limit the range of available prediction options and increase coding costs at low bitrates.

Innovation Solution

A method that detects redundant spatial prediction modes and generates additional modes by enlarging the neighboring pixel zones, replacing redundant modes with new prediction possibilities, thereby enriching the range of available predictions and improving coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple spatial prediction modes are used in H.264/AVC standard, then prediction coverage is improved, but redundant modes increase coding cost and reduce efficiency

Engineering Contradiction:
Improveprediction mode coverageVSAvoidcoding cost
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts and removes redundant prediction modes from the set of available spatial prediction modes. By identifying modes that generate similar prediction blocks and eliminating them, the system reduces the number of modes that need to be coded and transmitted, thereby reducing coding cost while maintaining prediction coverage through the remaining non-redundant modes.

Inventive Principle:
Principle #2Taking out (Extraction)

2Adaptability or versatility

If redundant prediction modes are maintained, then prediction options are available, but coding efficiency decreases due to unnecessary mode selection

Engineering Contradiction:
Improveprediction mode optionsVSAvoidcoding efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent removes redundant prediction modes from the available options, extracting only the necessary non-redundant modes. This extraction process eliminates unnecessary mode selection overhead and improves coding efficiency by reducing the search space and decision complexity, while maintaining adequate prediction options through the retained non-redundant modes.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent discards redundant prediction modes that provide duplicate or similar prediction capabilities. By eliminating these redundant options, the system recovers coding efficiency without significantly compromising prediction quality, as the remaining non-redundant modes suffice to represent the essential prediction directions and patterns.

Inventive Principle:
Principle #34Discarding and recovering

3Reliability

If all spatial prediction modes are coded in bitstream, then complete prediction information is transmitted, but data volume increases

Engineering Contradiction:
Improveprediction information completenessVSAvoidbitstream volume
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts and eliminates redundant prediction modes from the bitstream encoding process. By identifying and removing modes that generate similar prediction blocks, the system reduces the number of mode indices that need to be transmitted, thereby reducing bitstream volume while maintaining reliable prediction information through the remaining non-redundant modes.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP2870753B1Spatial prediction with increased number of possible coding modes
Publication Date: 2018.04.25 THOMSON LICENSING SA
  • EP2870753B1 patent drawingFigure 1
  • EP2870753B1 patent drawingFigure 2
  • EP2870753B1 patent drawingFigure 3

AI summary

The invention relates to a spatial prediction method of a block of pixels of an image, called current block, using a plurality of spatial prediction modes based on neighbouring pixels of the current block, comprising a step for determining (E3) a prediction block of the current block according to each one of said spatial prediction modes and a step for selecting (E4) one of said prediction blocks according to a predetermined criterion. According to the invention, before carrying out the preceding steps, it is detected (E1), among said spatial prediction modes, whether some of them are redundant. If such redundant modes are detected, one of them is replaced (E2) by an additional mode, distinct from the other spatial prediction modes