Intra Prediction Mode Derivation for Video Block Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video compression standards, such as HEVC, face high coding costs when neighboring blocks are not coded using intra prediction modes, and the introduction of new prediction modes affects subsequent blocks, necessitating an improved method for image or video data compression.

Innovation Solution

A method that determines and associates an intra prediction mode with a block based on neighboring blocks, even if they are not coded using intra prediction, allowing for the use of fictitious modes to predict subsequent blocks, and selects the best prediction mode to minimize distortion, thereby optimizing coding and decoding processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the MPM list construction mechanism is used when neighboring blocks are not coded using intra prediction modes, then the coding process can proceed, but the coding cost becomes high

Engineering Contradiction:
Improvecoding efficiencyVSAvoidcoding cost
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent makes the decoding process self-sufficient by determining intra prediction modes for neighboring blocks even when they are not coded using intra prediction modes. This is achieved by deriving prediction modes from spatial or temporal relationships with other blocks, eliminating the need for costly default mode assignments and reducing dependency on external information.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent performs preliminary determination of intra prediction modes for neighboring blocks before the actual decoding of the current block. By pre-establishing the prediction mode information for neighboring blocks through spatial or temporal derivation, the system prepares the necessary data in advance, avoiding the need for expensive default mode assignments during the main decoding process.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If new intra prediction modes are introduced in addition to existing classic intra prediction modes, then prediction accuracy may improve, but the cost of coding intra prediction modes for subsequent blocks increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidcoding cost for subsequent blocks
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a unified prediction mode determination mechanism that works for both classic intra prediction modes and new intra prediction modes. The same spatial and temporal derivation processes are used regardless of the specific mode type, allowing the system to handle multiple prediction mode kinds through a single universal framework, thereby avoiding increased coding costs for subsequent blocks.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent changes the fundamental parameter of how prediction modes are selected - instead of using separate mechanisms for different mode types, it applies a consistent parameter-based approach where prediction modes are derived from spatial relationships with neighboring blocks or temporal relationships with previous blocks. This parameter change ensures that introducing new modes does not inherently increase coding complexity.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4468705A1Methods and devices for encoding and decoding a data stream representative of at least one image
Publication Date: 2024.11.27 ORANGE SA
  • EP4468705A1 patent drawingFigure 1
  • EP4468705A1 patent drawingFigure 2~5
  • EP4468705A1 patent drawingFigure 4

AI summary

The invention relates to a method for encoding and a method for decoding a stream of encoded data representing at least one image divided into blocks. For at least one block of the image, called the current block, it is determined (E42) whether the current block is encoded according to an intra-block coding mode or another coding mode, the intra-block coding mode being a coding mode using an intra-block prediction mode selected from a group of intra-block prediction modes, based on at least one intra-block prediction mode associated with a block adjacent to the current block, the other coding mode being a coding mode in which the prediction of the current block is obtained by encoding the block pixel by pixel according to a local prediction function using previously reconstructed pixels adjacent to the current pixel.When the current block is encoded according to said intra-encoding mode, an intra-encoding mode from said intra-encoding mode group is determined (E431) for the current block, based on at least one intra-encoding mode associated with a previously decoded block of the image. The current block is decoded (E43) based on said determined intra-encoding mode that is associated (E437) with said current block. When the current block is encoded according to said other encoding mode, the current block is decoded (E44) according to said other encoding mode. An intra-encoding mode from said intra-encoding mode group is determined (E449) and associated (E450) with said current block. When decoding a block following the current block, said intra-encoding mode previously associated with the current block for predicting an intra-encoding mode of the following block is used if said following block is encoded according to an intra-encoding mode.