Video Coding Geometric Partitioning Mode Refinement

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional video coding techniques suffer from low coding efficiency, which is undesirable for digital video applications.

Innovation Solution

A method of processing video data that involves obtaining a geometric partitioning mode (GPM) block associated with a current video block during conversion to a bitstream, and performing the conversion using a motion-compensated prediction sample refinement process applied to the GPM block.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If conventional video coding techniques are used, then device complexity is reduced, but coding efficiency deteriorates

Engineering Contradiction:

Improvecoding efficiencyVSAvoidprocessing complexity

Core Design Contradiction:

Loss of energyVSDevice complexity

Solution Approach 1:

The current video block is divided into two triangular partitions using geometric partitioning mode (GPM), where each triangle is further subdivided into sub-blocks. This segmentation allows independent motion compensation and prediction sample refinement to be applied to each partition, improving coding efficiency by capturing different motion characteristics in different regions of the block.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Motion-compensated prediction sample refinement is applied selectively to GPM blocks based on specific conditions (e.g., block size, motion complexity). The refinement process adjusts prediction samples locally within each triangular partition using gradient-based optimization, improving prediction accuracy where needed while avoiding unnecessary processing in simpler regions.

Inventive Principle:

Principle #3Local quality

2Loss of energy

If motion-compensated prediction sample refinement is applied to GPM blocks, then coding efficiency is improved, but processing time increases

Engineering Contradiction:

Improvecoding efficiencyVSAvoidprocessing time

Core Design Contradiction:

Loss of energyVSLoss of time

Solution Approach 1:

Motion compensation is performed first to generate initial prediction samples for each triangular partition before applying sample refinement. The refinement process then adjusts these pre-computed samples using gradient calculations, avoiding the need to perform full optimization from scratch and reducing overall processing time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The sample refinement is applied selectively rather than universally to all video blocks. The patent applies refinement only to GPM blocks that meet specific criteria (such as block size thresholds or motion complexity measures), performing partial refinement where beneficial while skipping processing for blocks where refinement would not significantly improve coding efficiency.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250080718A1Method, device, and medium for video processing

Publication Date: 2025.03.06 DOUYIN VISION CO LTD

AI summary

Embodiments of the present disclosure provide a solution for video processing. A method of processing video data is proposed. The method comprises: during a conversion between a current video block of a video and a bitstream of the video, obtaining a geometric partitioning mode (GPM) block associated with the current video block; and performing the conversion based on a motion-compensated prediction sample refinement process applied to the GPM block. Compared with the conventional solution, the proposed solution advantageously improves coding efficiency and compression ratios.