Piecewise Chroma-Luma Prediction for Lower Video Bitrate
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codecs face challenges in efficiently encoding and decoding video data, particularly in managing cross-component predictions between different color channels, leading to increased complexity and bitrate due to non-linearities and discontinuities in modeling dependencies.
Innovation Solution
A method and apparatus that utilize two cross-component models, determined by threshold values, to predict chroma values based on luma values, using average or median representatives, and applying different models based on the input sample's value relative to the threshold, reducing complexity and bitrate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single cross-component model is used to predict chroma values from luma values, then the encoding process is simple, but discontinuities and non-linearities occur leading to increased complexity and bitrate
Solution Approach 1:
The patent divides the cross-component prediction into multiple linear models, each valid for a specific range of luma values. By segmenting the prediction space and applying different linear models to different segments, the method avoids the discontinuities and non-linearities that would occur with a single model, thereby reducing complexity while maintaining prediction accuracy.
2Reliability
If multiple cross-component models are used to reduce discontinuities, then prediction accuracy improves, but the encoding complexity and bitrate increase
Solution Approach 1:
The patent applies different linear models with locally optimized parameters to different regions of the luma value range. Each model is tailored to the specific characteristics of its local region, improving prediction accuracy where needed while keeping the overall system manageable through localized rather than global complexity.
3Loss of information
If cross-component prediction is applied to all input samples, then encoding completeness is maintained, but bitrate increases due to transmitting model parameters for all samples
Solution Approach 1:
The patent segments the input samples based on their luma values and applies different linear models only to relevant segments. This selective application reduces the number of samples requiring full model parameter transmission, thereby reducing bitrate while maintaining encoding completeness for all samples through the segmented approach.
Data Source
AI summary
The embodiments relate to a method for processing image and/or video data. The method comprises receiving a set of input samples; determining a set of reference samples with two types of colour information; determining a center representative of both of said two types of colour information; determining a lower representative and a higher representative based on the set of reference samples and the determined center representative of at least one of the two types of colour information; determining a first cross-component model according to the determined lower and center representatives, and determining a second cross-component model according to the determined higher and center representatives; and determining when a value of an input sample is smaller than or equal to at least one of the determined center representatives, and applying the first cross-component model to the input sample, otherwise applying the second cross-component model to the input sample.