Variable-Size Sub-Block Motion Prediction for Video Coding

Solution Overview

Problem

Existing video coding technologies struggle to accurately represent camera and object motions beyond simple translation, such as zoom, rotation, and perspective, leading to inefficiencies in video compression.

Innovation Solution

Implementing sub-block based motion prediction techniques, including affine transform, alternative temporal motion vector prediction, spatial-temporal motion vector prediction, bi-directional optical flow, and frame-rate up conversion, to enhance motion compensation by dividing video blocks into non-uniform and variable sub-block sizes for more accurate motion information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If uniform sub-block sizes are used for motion prediction, then the processing is simple and efficient, but the motion compensation accuracy is insufficient for complex camera and object motions

Engineering Contradiction:

Improvemotion compensation accuracyVSAvoidsub-block partitioning complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The video block is divided into multiple sub-blocks with different sizes based on the motion characteristics of each region. This segmentation allows each sub-block to be processed with appropriate motion prediction techniques, improving overall motion compensation accuracy while managing complexity through adaptive partitioning.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different sub-blocks within the same video block are assigned different sizes and motion prediction methods according to their local motion characteristics. Regions with complex motion receive more detailed processing with smaller sub-blocks, while regions with simple motion use larger sub-blocks, optimizing the balance between accuracy and complexity.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If more sophisticated motion prediction algorithms are applied, then motion compensation accuracy improves, but processing time and computational complexity increase

Engineering Contradiction:

Improvemotion prediction accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The motion prediction system dynamically selects and switches between different prediction algorithms based on the motion characteristics detected in each sub-block. This allows the system to use computationally intensive algorithms only where necessary, improving accuracy for complex motions while maintaining efficiency for simple motions.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

Instead of applying complex motion prediction algorithms uniformly to the entire block, the system applies sophisticated algorithms only to sub-blocks that require them based on detected motion characteristics. This partial application reduces overall processing time while maintaining high accuracy where needed.

Inventive Principle:

Principle #16Partial or excessive action

3Measurement precision

If sub-block sizes are made variable and non-uniform, then motion information representation accuracy improves, but the complexity of processing and encoding increases

Engineering Contradiction:

Improvemotion information accuracyVSAvoidprocessing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The video block is divided into multiple sub-blocks with different sizes based on the motion characteristics of each region. This segmentation allows each sub-block to be processed with appropriate motion prediction techniques, improving overall motion compensation accuracy while managing complexity through adaptive partitioning.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different sub-blocks within the same video block are assigned different sizes and motion prediction methods according to their local motion characteristics. Regions with complex motion receive more detailed processing with smaller sub-blocks, while regions with simple motion use larger sub-blocks, optimizing the balance between accuracy and complexity.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250294179A1Concept Of Interweaved Prediction

Publication Date: 2025.09.18 DOUYIN VISION CO LTD

AI summary

Methods, systems, and devices related to sub-block based motion prediction in video coding are described. In one representative aspect, a video processing method includes partitioning a video block into a first set of sub-blocks according to a first pattern, partitioning the video block into a second set of sub-blocks according to a second pattern, in which at least one sub-block in the second set has a different size than a sub-block in the first set, and determining a prediction block corresponding to a combination of a first intermediate prediction block that is predictively generated from the first set of sub-blocks and a second intermediate prediction block that is predictively generated from the second set of sub-blocks.