Cb/Cr Chroma Mode Coding with Component-Specific CCLM

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional video coding techniques fail to distinguish between the different properties of the Cb and Cr channels in chroma images, leading to suboptimal encoding and decoding performance and reduced video quality.

Innovation Solution

Adopting a method that encodes and decodes component-specific intra-prediction modes for each chroma channel (Cb and Cr) using cross-component linear model flags and indices to adapt to the unique features of each channel.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If the same intra-prediction mode is applied to both Cb and Cr channels, then the encoding process is simplified, but the prediction accuracy and video quality deteriorate due to not accounting for channel-specific properties

Engineering Contradiction:
Improveencoding process complexityVSAvoidprediction accuracy
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent segments the chroma prediction process by separately encoding intra-prediction mode information for the Cb and Cr channels. Instead of using a single shared mode, the system divides the prediction mode selection into channel-specific processes, allowing each channel to have its own optimal prediction mode based on its unique statistical properties and visual characteristics.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by allowing different prediction modes for different chroma channels. The Cb channel and Cr channel are treated with different prediction strategies tailored to their specific properties, improving the overall prediction accuracy by adapting the prediction quality to local channel characteristics rather than applying a uniform approach.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If component-specific intra-prediction modes are encoded for each chroma channel, then video coding efficiency and quality improve, but the bitstream complexity and encoding overhead increase

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

Solution Approach 1:

The patent implements dynamics by making the prediction mode selection adaptive and flexible. The encoder dynamically chooses between different prediction modes (angular, planar, DC) based on the specific characteristics of each chroma channel and the current block properties. This dynamic adaptation allows the system to optimize quality while managing complexity through intelligent mode selection rather than fixed complex structures.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes parameters by introducing separate prediction mode parameters for Cb and Cr channels. Instead of using a single prediction mode parameter, the system now manages multiple parameters (Cb prediction mode, Cr prediction mode, and their respective mode flags), allowing independent optimization of each channel's prediction characteristics to achieve better quality.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If more intra-prediction modes are introduced to handle higher spatial resolutions, then prediction accuracy improves, but the encoding complexity and processing time increase

Engineering Contradiction:
Improveprediction accuracy at high resolutionVSAvoidencoding processing speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-defining the set of available prediction modes and their corresponding parameters before the encoding process. The system prepares the prediction mode options in advance and uses efficient parameter encoding techniques to select from these pre-defined modes, reducing the computational burden during actual encoding while maintaining access to multiple accurate prediction modes for high-resolution content.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20250310539A1Method and apparatus for encoding intra prediction mode for each chroma component
Publication Date: 2025.10.02 HYUNDAI MOTOR CO LTD
  • US20250310539A1 patent drawing
  • US20250310539A1 patent drawing
  • US20250310539A1 patent drawing

AI summary

A method and an apparatus are disclosed for coding an intra-prediction mode for each chroma component. In the disclosed embodiments, a video decoding device decodes, from a bitstream, first intra-prediction mode information for a first chroma channel of the current chroma block. The video decoding device decodes, from the bitstream, second intra-prediction mode information for a second chroma channel of the current chroma block. Here, the first intra-prediction mode information includes at least one of a first CCLM (cross-component linear model) mode flag, a first CCLM mode index, and a first chroma intra-prediction mode indicator, and the second intra-prediction mode information includes at least one of a second CCLM mode flag, a second CCLM mode index, and a second chroma intra-prediction mode indicator.