Cross-Component Prediction Filter Selection for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques struggle to achieve optimal coding efficiency, particularly in predicting chroma sample values from luma samples, due to limitations in cross-component prediction tools.
Innovation Solution
The proposed method involves a cross-component prediction technique that selects a filter shape candidate based on external luma and chroma sample values to predict internal chroma sample values, using a set of weighting coefficients that include those for non-linear values derived from scaled products of luma samples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional cross-component prediction tools are used to predict chroma sample values from luma samples, then video coding can be performed, but coding efficiency is suboptimal due to limitations in prediction accuracy
Solution Approach 1:
The patent changes the parameters of the prediction model by introducing multiple filter shape candidates with different weighting coefficients. Instead of using a fixed prediction model, the system evaluates multiple candidates with varying parameters (filter shapes, weighting coefficients) to select the one that best matches the current video content characteristics, thereby improving prediction accuracy and coding efficiency
Solution Approach 2:
The patent makes the prediction model dynamic by selecting filter shape candidates adaptively based on external luma and chroma sample values. The system dynamically adjusts which filter shape candidate is used for prediction, allowing the prediction accuracy to adapt to different video content characteristics, thus resolving the contradiction between prediction accuracy and coding efficiency
2Reliability
If more complex prediction models are used to improve chroma prediction accuracy, then video quality is maintained, but computational complexity and processing overhead increase
Solution Approach 1:
The patent segments the prediction process into distinct stages: obtaining external sample values, evaluating multiple filter shape candidates, selecting the best candidate, and performing the actual prediction. This segmentation allows the system to manage complexity by breaking down the complex prediction task into manageable steps, maintaining video quality while controlling processing complexity
Solution Approach 2:
The patent applies partial action by using external luma and chroma sample values from a limited external region rather than the entire video block. This partial use of available data reduces computational complexity while still providing sufficient information for accurate prediction, thus maintaining video quality without excessive processing requirements
Data Source
AI summary
The present disclosure provides a method for decoding video data, comprising: obtaining a video block from a bitstream; obtaining internal luma sample values of the video block, external luma sample values of an external region of the video block and external chroma sample values of the external region; selecting, based on the external luma sample values and the external chroma sample values, one filter shape candidate from a plurality of filter shape candidates for predicting internal chroma sample values of the video block, wherein the plurality of filter shape candidates correspond to different sets of weighting coefficients for predicting chroma sample values based on corresponding luma sample values; predicting, with the selected filter shape candidate, the internal chroma sample values based on the internal luma sample values; and obtaining decoded video block using the predicted internal chroma sample values.