CfL Downsampling Filter Signaling for Chroma Intra Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding technologies, such as AV1 and HEVC, lack efficient methods for signaling downsampling filters in cross-component intra prediction modes, particularly for chroma from luma, which can lead to suboptimal performance in video compression.

Innovation Solution

A method and device for performing chroma from luma (CfL) intra prediction that involves receiving a current chroma block, determining if CfL intra prediction mode is enabled, obtaining a syntax element for the downsampling filter, selecting an appropriate filter, and reconstructing the chroma block using downsampled luma samples, along with a non-transitory computer-readable medium storing instructions for processors to perform these operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If downsampling filters are not signaled for chroma from luma intra prediction mode, then device complexity is reduced, but video compression efficiency deteriorates

Engineering Contradiction:

Improvefilter signaling complexityVSAvoidvideo compression efficiency

Core Design Contradiction:

Device complexityVSProductivity

Solution Approach 1:

The patent changes the parameter being signaled from the full filter kernel to a simplified filter index or type indicator. Instead of transmitting complete filter coefficients, the encoder signals a compact representation that identifies which filter from a predefined set should be used, reducing the bit cost while maintaining filter functionality

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

Instead of having the decoder determine the filter from complex syntax elements, the patent inverts the approach by having the encoder explicitly signal a compact filter identifier. This reversal simplifies the decoding process and reduces the computational burden on the decoder while preserving compression efficiency

Inventive Principle:

Principle #13The other way round (Inversion)

2Adaptability or versatility

If multiple downsampling filters are supported for different color formats, then adaptability is improved, but device complexity increases

Engineering Contradiction:

Improvecolor format compatibilityVSAvoidfilter selection complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies different filter characteristics tailored to specific color formats (e.g., 4:2:0, 4:2:2, 4:4:4). Each color format has its own optimized filter from the predefined set, allowing the system to adapt locally to the specific requirements of each format without implementing a single complex universal filter

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

A single unified filter set is designed that can handle multiple color formats through a common framework. The same filter structure is reused across different formats with minor parameter adjustments, reducing overall device complexity compared to implementing separate filters for each format

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Measurement precision

If downsampling filter optimization is implemented for chroma blocks, then prediction accuracy is improved, but computational complexity increases

Engineering Contradiction:

Improvechroma prediction accuracyVSAvoiddownsampling computation

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies downsampling filters selectively based on block size, color format, and prediction mode rather than uniformly to all chroma blocks. This partial application approach maintains high prediction accuracy for blocks where it matters most while reducing computational overhead for blocks where simple averaging suffices

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The downsampling filters are pre-computed and stored in lookup tables during encoder initialization. Instead of computing filters in real-time during decoding, the pre-computed filter kernels are readily available, significantly reducing the computational complexity during actual video decoding while maintaining prediction accuracy

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260006192A1Signaling of downsampling filters for chroma from luma intra prediction mode

Publication Date: 2026.01.01 TENCENT AMERICA LLC

AI summary

An example method of video coding includes receiving a video bitstream comprising a plurality of blocks including a chroma block. When a chroma-from-luma (CfL) intra prediction mode is enabled for the chroma block, a syntax element is parsed from the video bitstream, the syntax element indicating which downsampling filter to use for the chroma block. The method also includes selecting a downsampling filter from a plurality of downsampling filters according to the syntax element, where the plurality of downsampling filters include a 6-tap downsampling filter and a 4-tap downsampling filter. The method further includes determining luma sample positions associated with the chroma block based on the selected downsampling filter, and downsampling a plurality of luma samples at the luma sample positions, wherein pixels in each downsampled luma sample are co-located with corresponding pixels in the chroma block.