Reinforcement-Learning Video Coding for 4K/8K Block Partitioning

Solution Overview

Problem

Existing video coding methods, such as those based on the Versatile Video Coding (VVC) standard, struggle to adapt to high-definition videos like 4K or 8K due to increased calculation complexity, leading to low efficiency and timeliness issues in video coding.

Innovation Solution

Implement reinforcement learning network training to optimize coding unit partitioning, combining backward and forward neural network training with fusion to determine optimized coding unit partitioning information, and evaluate coding contents using a preset evaluation index to identify a suitable coding mode.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional recursive block partitioning is used in VVC standard, then video compression efficiency is improved, but calculation complexity increases and processing time is prolonged

Engineering Contradiction:

Improvevideo compression efficiencyVSAvoidcalculation complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent performs preliminary classification of video blocks into different types (e.g., homogeneous, heterogeneous, edge-containing blocks) before applying partitioning. This preliminary action allows the system to select appropriate partitioning strategies in advance, avoiding exhaustive recursive traversal of all possible partitioning modes, thereby reducing calculation complexity while maintaining compression efficiency

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent dynamically adjusts partitioning parameters such as block size, partitioning depth, and partitioning mode based on local video content characteristics (e.g., texture complexity, motion intensity). By changing these parameters adaptively rather than using fixed recursive partitioning, the system achieves good compression efficiency with reduced computational burden

Inventive Principle:

Principle #35Parameter changes

2Productivity

If traditional recursive block partitioning is used in VVC standard, then video compression efficiency is improved, but video coding timeliness deteriorates

Engineering Contradiction:

Improvevideo compression efficiencyVSAvoidprocessing time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent performs preliminary classification of video blocks into different types (e.g., homogeneous, heterogeneous, edge-containing blocks) before applying partitioning. This preliminary action allows the system to select appropriate partitioning strategies in advance, avoiding exhaustive recursive traversal of all possible partitioning modes, thereby reducing calculation complexity while maintaining compression efficiency

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent dynamically adjusts partitioning parameters such as block size, partitioning depth, and partitioning mode based on local video content characteristics (e.g., texture complexity, motion intensity). By changing these parameters adaptively rather than using fixed recursive partitioning, the system achieves good compression efficiency with reduced computational burden

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If existing compression methods are used, then current video definition requirements are met, but adaptability to higher definitions (4K, 8K) is insufficient

Engineering Contradiction:

Improveadaptability to video definitionVSAvoidvideo coding efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent implements dynamic partitioning strategies that adapt to different video definitions (HD, 4K, 8K) by adjusting block sizes, partitioning depths, and processing granularities based on the resolution and content characteristics. This dynamic approach enables the same compression framework to efficiently handle various definitions without sacrificing coding efficiency

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent dynamically adjusts partitioning parameters such as block size, partitioning depth, and partitioning mode based on local video content characteristics (e.g., texture complexity, motion intensity). By changing these parameters adaptively rather than using fixed recursive partitioning, the system achieves good compression efficiency with reduced computational burden

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250260817A1Video coding method, network device, and storage medium

Publication Date: 2025.08.14 SANECHIPS TECH CO LTD

AI summary

A video coding method, a network device, and a computer-readable storage medium are disclosed. The method includes: performing reinforcement learning network training on an acquired image sequence to obtain optimized coding unit partitioning information corresponding to the image sequence; obtaining game result information corresponding to optimized coding contents according to the optimized coding contents, reference coding contents, and a preset evaluation index, with the optimized coding contents being obtained according to the optimized coding unit partitioning information, and the reference coding contents being obtained according to the image sequence; and when it is determined that the game result information meets a preset game condition, determining a coding mode corresponding to the optimized coding unit partitioning information as a video coding mode for the image sequence.