Distance-Based Interpolation Filtering for Intra-Frame Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intra-frame prediction interpolation filters in AVS 3 use the same group of filter coefficients for pixels in different locations, leading to inaccurate predicted values and affecting encoding and decoding efficiency, while smoothing effects result in loss of high-frequency texture detail information.
Innovation Solution
Adaptively select interpolation filter parameters based on the location information of prediction pixels within a specified prediction unit, using distance-based criteria to choose appropriate filters for improved accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the same group of filter coefficients is used for pixels in different locations, then the device complexity is reduced, but the measurement precision of pixel interpolation prediction deteriorates
Solution Approach 1:
The patent applies local quality by differentiating filter coefficient selection based on pixel location. Different filter coefficients are selected according to the position of prediction pixels within the prediction unit, allowing each region to receive optimized filtering parameters that match its specific characteristics, thereby improving prediction accuracy without uniformly increasing complexity across the entire image.
Solution Approach 2:
The patent implements dynamics by making the filter coefficient selection adaptive rather than static. The system dynamically selects appropriate filter coefficients based on the location information of prediction pixels, allowing the filtering parameters to change according to the specific requirements of different regions, thus optimizing the balance between complexity and precision.
2Reliability
If smoothing effect is applied to alleviate noise impact, then the reliability of prediction is improved, but the loss of high-frequency texture detail information increases
Solution Approach 1:
The patent applies local quality by selecting different filter coefficients based on pixel location, which allows the smoothing strength to be locally adjusted. Regions with more noise can receive stronger smoothing, while regions with important high-frequency details can receive weaker smoothing or different filtering characteristics, thus balancing noise reduction with detail preservation.
Solution Approach 2:
The patent changes the filtering parameters (filter coefficients) based on the location of prediction pixels. By varying these parameters according to spatial position, the system can optimize the trade-off between smoothing and detail preservation for different regions, preventing uniform over-smoothing that would lose high-frequency information.
3Reliability
If excessively smooth interpolation filter is used, then the reliability of prediction is improved, but the encoding performance gain is reduced due to excessively small difference between predicted values from adjacent prediction angles
Solution Approach 1:
The patent applies local quality by selecting filter coefficients based on pixel location within the prediction unit. This allows different regions to have different smoothing strengths, ensuring that prediction accuracy is maintained where needed while preserving sufficient variation between adjacent prediction angles for optimal encoding performance.
Solution Approach 2:
The patent changes the filter parameter selection based on location information. By dynamically adjusting which filter coefficients are applied according to pixel position, the system optimizes the balance between prediction reliability and the diversity of predicted values from different prediction angles, thereby maintaining encoding efficiency.
Data Source
AI summary
The embodiments of the disclosure provide an interpolation filtering method and apparatus for intra-frame prediction, a medium, and an electronic device. According to a distance between a prediction pixel and a first reference pixel, a target filter matching the distance is selected from a filter parameter corresponding to a specified prediction unit, and interpolation filtering processing is performed on a second reference pixel by using the target filter, to obtain a predicted value of the prediction pixel.