Boundary Partition Syntax for Video Decoding Complexity Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video coding standards face challenges in balancing bandwidth requirements and video quality, particularly with higher resolution videos that require more bandwidth, leading to increased coding complexity without significant quality improvement.

Innovation Solution

The proposed method employs a versatile boundary partition technique that uses a consistent block partition syntax for both boundary and non-boundary blocks, omitting inverse transform operations on boundary blocks when not split, thereby reducing decoding complexity and maintaining continuity in context-adaptive binary arithmetic coding (CABAC) engines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If higher resolution video is used to improve video quality, then video quality is improved, but bandwidth requirements and coding complexity increase

Engineering Contradiction:
Improvevideo qualityVSAvoidcoding complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The video picture is divided into multiple blocks that can be independently processed. Boundary blocks are identified and handled separately from non-boundary blocks, allowing selective application of inverse transform operations only where necessary, thereby reducing overall coding complexity while maintaining video quality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different processing strategies are applied to different regions of the video picture. Boundary blocks receive simplified processing (omitting inverse transform) while non-boundary blocks receive full processing, optimizing the balance between quality and complexity locally across the picture.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If inverse transform operations are performed on all blocks to maintain video quality, then video quality is maintained, but decoding complexity increases

Engineering Contradiction:
Improvevideo qualityVSAvoiddecoding complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The inverse transform operation is extracted and selectively applied only to non-boundary blocks. Boundary blocks are identified and excluded from this operation, removing unnecessary processing steps while preserving video quality in regions where the operation would have minimal impact.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of applying inverse transform to all blocks (excessive action), the method applies it only to non-boundary blocks (partial action), which is sufficient to maintain video quality while significantly reducing decoding complexity.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If different partition syntax is used for boundary and non-boundary blocks to optimize processing, then processing efficiency is improved, but CABAC engine continuity is broken

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidCABAC engine continuity
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

A single unified partition syntax is used for both boundary and non-boundary blocks, making the syntax universal. This single syntax serves multiple functions: it correctly identifies boundary blocks for simplified processing while maintaining continuity for the CABAC engine, eliminating the need for separate syntax sets.

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

Data Source

PatentUS11539953B2Apparatus and method for boundary partition
Publication Date: 2022.12.27 HUAWEI TECH CO LTD
  • US11539953B2 patent drawing
  • US11539953B2 patent drawing
  • US11539953B2 patent drawing

AI summary

A method for boundary partition of a current block is provided. The method includes obtaining a partition indicator from a received bitstream; determining whether the partition indicator indicates that the current block is to be split and whether the current block is a boundary block; and performing a non-inverse transform operation on an inside part of the current block when the partition indicator indicates that the current block is not to be split and the current block is a boundary block, wherein the inside part of the block is located within the picture. For the boundary block which is determined not to be split, a non-inverse transform operation is chosen. This avoids a normal inverse transform on the residuals of the inside part (within the picture) of the current block in the decoder, so number of blocks for the inverse transformation is decreased, and decoding complexity is thus reduced.