Video Coding Chroma Prediction Using Mixed Autocorrelation Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods, such as Versatile Video Coding (VVC/H.266), fail to effectively utilize non-local correlations and large-scale trends in chroma and luma channels, leading to inefficient data compression due to the lack of reliable texture models during decoding.
Innovation Solution
The method involves reconstructing samples in luminance and chrominance channels, determining a reference area for prediction, computing autocorrelation matrices, mixing these matrices with neighboring blocks, and applying filter coefficients to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional prediction methods (CCLM/CCCM) are used based on neighboring samples, then decoding complexity is kept low, but prediction accuracy is insufficient due to lack of non-local correlation utilization
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing autocorrelation matrices during the encoding phase. These matrices capture statistical properties of the video content in advance, allowing the decoder to utilize pre-prepared statistical information without performing complex computations during decoding, thus improving prediction accuracy while maintaining low decoding complexity
Solution Approach 2:
The patent introduces autocorrelation matrices as an intermediary between the encoded video data and the prediction process. These matrices serve as a statistical bridge that captures non-local correlations and enables more accurate prediction without requiring the decoder to directly analyze complex spatial relationships in the video content
2Measurement precision
If more reference samples and larger reference areas are used for prediction, then prediction accuracy improves, but memory requirements and processing time increase
Solution Approach 1:
The patent pre-computes autocorrelation matrices during encoding and stores them for later use. This preliminary computation captures the statistical properties of large reference areas without requiring the decoder to process these extensive datasets in real-time, thus improving prediction accuracy while avoiding increased processing time during decoding
Solution Approach 2:
The patent replaces the mechanical process of directly analyzing and processing large numbers of reference samples during decoding with a statistical approach using pre-computed autocorrelation matrices. This substitution transforms the computationally intensive sample-by-sample analysis into efficient matrix-based statistical operations
3Productivity
If conventional linear models are used for chroma prediction, then implementation is simple, but compression efficiency is limited due to insufficient utilization of non-local correlations
Solution Approach 1:
The patent introduces autocorrelation matrices as a statistical intermediary that captures non-local correlations between luma and chroma components. These matrices enable the model to utilize long-range dependencies and complex relationships without requiring a complex computational structure, thus improving compression efficiency while maintaining reasonable model complexity
Solution Approach 2:
The patent changes the parameters used in the linear model by incorporating autocorrelation matrix information. Instead of relying solely on direct sample values, the model uses statistically-derived parameters from the autocorrelation matrices, enabling it to capture non-local correlations and improve compression efficiency
Data Source
AI summary
A method comprising: receiving an image block unit of a frame, the image block unit comprising samples in color channels comprising at least one chrominance channel and one luminance channel (1100); reconstructing samples of said luminance channels of the image block unit (1102); determining a reference area for predicting target samples of at least one color channel of the image block unit, wherein said reference area comprises one or more of reference samples in a neighboring block in current color channel/frame, in the neighboring of a co-located block in reference color channel/frame; and/or inside the co-located block in reference color channel/frame (1104); determining filter coefficients of a filter for said predicting based the reference samples and a shape of the filter (1106); reconstructing the target samples of at least one color channel of the image block unit using said prediction based on the samples of said reference area and the filter coefficients (1108); computing an autocorrelation matrix for the reconstructed target samples of the image block unit (1110); computing autocorrelation matrices for a plurality of neighboring image blocks units of said image block unit (1112); mixing the autocorrelation matrices of each of the plurality of neighboring image blocks units separately with the autocorrelation matrix of said image block unit (1114); and storing the mixed autocorrelation matrices for said image block unit (1116).