Video Block Mode Harmonization for CIIP Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding standards face challenges in efficiently utilizing combined intra-inter prediction to enhance video quality and reduce bandwidth requirements, particularly in advanced coding technologies like HEVC and future standards such as VVC.

Innovation Solution

Implementing combined intra-inter prediction (CIIP) techniques, including methods like history-based motion vector prediction (HMVP), symmetric motion vector difference (SMVD), and weighted averaging in video encoding and decoding processes to optimize motion vector prediction and improve coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If combined intra-inter prediction (CIIP) is implemented to improve video quality and reduce bandwidth, then coding efficiency is improved, but device complexity increases due to additional processing requirements

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprocessing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The video block is divided into two prediction units: one for intra prediction and one for inter prediction. This segmentation allows the decoder to process different prediction modes independently and combine them, improving coding efficiency while managing complexity through modular processing

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent combines intra prediction and inter prediction results through weighted averaging to produce the final reconstructed block. This merging of two prediction approaches leverages the strengths of both methods, achieving superior coding efficiency compared to using either method alone

Inventive Principle:
Principle #5Merging (Combining)

2Manufacturing precision

If multiple prediction modes (pairwise prediction, combined-bi prediction, CIIP) are supported to enhance video processing, then video quality improves, but ease of operation deteriorates due to increased mode selection complexity

Engineering Contradiction:
Improvevideo qualityVSAvoidmode selection complexity
Core Design Contradiction:
Manufacturing precisionVSEase of operation

Solution Approach 1:

The prediction mode is dynamically selected based on the coded mode of neighboring or reference blocks. The system adapts between pairwise prediction, combined-bi prediction, and CIIP modes according to the specific video content and block characteristics, improving video quality while automating mode selection to maintain ease of operation

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses feedback from the coded modes of neighboring and reference blocks to determine the appropriate prediction mode for the current block. This feedback mechanism allows automatic adaptation to local video characteristics, achieving high video quality without manual intervention

Inventive Principle:
Principle #23Feedback

3Productivity

If constrained flags are used to control prediction mode selection based on neighboring blocks, then coding efficiency improves, but device complexity increases due to additional constraint checking

Engineering Contradiction:
Improvecoding efficiencyVSAvoidconstraint checking complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system performs preliminary determination of the coded mode for neighboring and reference blocks before processing the current block. This preliminary action allows the decoder to pre-establish which prediction modes and flags are applicable, improving coding efficiency through better mode selection while reducing runtime complexity

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12425578B2Harmonization between different video process mode
Publication Date: 2025.09.23 BYTEDANCE INC
  • US12425578B2 patent drawing
  • US12425578B2 patent drawing
  • US12425578B2 patent drawing

AI summary

The present application relates to harmonization between different video process mode. A method for processing video includes: determining, during a conversion between a first block in a video data and a bitstream representation of the first block, one or more coded modes of one or more second blocks; determining, based on the one or more coded modes of the one or more second blocks, a coded mode constraint of the first block; and performing, at least by applying the coded mode constraint of the first block, the conversion; wherein the one or more second blocks comprise at least one of an adjacent block, a non-adjacent block, and a reference block of the first block.