Component-Dependent Prediction Signaling for Video Compression

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face inefficiencies in intra and inter prediction methods, particularly in handling multiple color components, leading to suboptimal compression ratios and increased bit usage for less likely prediction directions and motion vectors.

Innovation Solution

The proposed method involves determining separate prediction modes for different color components within a block, using signaling information and context-based entropy coding to select the most efficient prediction mode, and employing advanced motion vector prediction techniques to reduce redundancy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If separate prediction modes are determined for different color components, then video coding efficiency is improved and compression ratios increase, but device complexity and processing overhead increase

Engineering Contradiction:
Improvevideo coding efficiencyVSAvoidprediction mode processing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the prediction mode determination process into separate handling for different color components (luma and chroma). Each color component can have its own prediction mode selected independently, allowing optimized processing for each component type while maintaining overall coding efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic selection mechanisms where prediction modes can be switched between intra and inter modes based on content characteristics. The decoder can adaptively determine whether to apply separate prediction modes for different color components, allowing flexibility to optimize for each specific video sequence or block type.

Inventive Principle:
Principle #15Dynamics

2Quantity of substance

If context-based entropy coding is used to select prediction modes, then bit usage is reduced for less likely directions, but encoding complexity increases

Engineering Contradiction:
Improvebit usageVSAvoidencoding complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent employs context-based entropy coding that dynamically adjusts probability models based on the current prediction context. The entropy coder adapts its parameters (probability estimates) according to the likelihood of different prediction modes and directions, allowing more efficient bit representation for statistically likely cases while maintaining reasonable coding for unlikely cases.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If advanced motion vector prediction techniques are employed, then redundancy is reduced and compression improves, but processing time and computational load increase

Engineering Contradiction:
Improvecompression ratioVSAvoidprocessing time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent implements motion vector prediction techniques that prepare and cache prediction information in advance. By pre-computing and storing motion vector candidates and prediction data, the decoder can quickly retrieve and apply these predictions during the decoding process, reducing real-time processing requirements while maintaining compression efficiency.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20250324036A1Component dependent intra and inter prediction signaling
Publication Date: 2025.10.16 TENCENT AMERICA LLC
  • US20250324036A1 patent drawing
  • US20250324036A1 patent drawing
  • US20250324036A1 patent drawing

AI summary

A non-transitory computer-readable medium storing a video media bitstream that is encoded by an encoding method, in which a first prediction mode for a first color component associated with a current block is determined. When the first prediction mode is determined as one of (i) an intra prediction mode, (ii) an inter prediction mode, or (iii) a combined intra-inter prediction mode, a second prediction mode for a second color component associated with the current block is determined based on the first prediction mode being determined from (i), (ii), or (iii). When the first prediction mode is the inter prediction mode, the second prediction mode is determined based on a template matching (TM) cost of a template of the second color component. The template includes adjacent samples of the second color component. The current block is encoded into the video media bitstream.