Synthesis Prediction Weighting for Image Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding and decoding techniques, such as those used in VVC, are insufficient in reducing code amount solely by averaging and synthesizing inter and intra prediction images.
Innovation Solution
The proposed technique involves generating a predicted image through synthesis prediction by weighting and averaging inter and intra prediction images, with the weighting parameter determined based on the type of intra prediction used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple averaging synthesis is used to combine inter and intra prediction images, then the processing complexity is reduced, but the code amount reduction effect is insufficient
Solution Approach 1:
The patent introduces weighting parameters to change the parameters of the synthesis process. Instead of simple averaging, the invention uses weighted averaging where the weights are determined based on the intra prediction type (e.g., planar, angular, DC). This parameter change enables better adaptation to different prediction scenarios, achieving superior code amount reduction while maintaining reasonable processing complexity.
2Loss of information
If weighting processing is applied to synthesize predicted images, then the code amount is reduced, but the processing complexity increases
Solution Approach 1:
The patent applies local quality by differentiating the weighting parameters based on the specific intra prediction type used for each block. Different intra prediction modes (planar, angular, DC) receive different weighting treatments. This localized adaptation optimizes the synthesis for each block's specific characteristics, achieving effective code amount reduction without requiring complex global processing.
Solution Approach 2:
The invention changes the weighting parameters dynamically based on the intra prediction type. By adjusting these parameters according to the prediction mode (planar, angular, DC), the system achieves optimized compression for each local region while maintaining a relatively simple overall processing framework.
3Device complexity
If a fixed weighting parameter is used for all intra prediction types, then the processing is simplified, but the image quality deteriorates
Solution Approach 1:
The patent implements local quality by assigning different weighting parameters to different intra prediction types. Instead of using a universal fixed weight, the system adapts the weighting locally based on whether the prediction is planar, angular, or DC. This localized optimization preserves image quality for each block type while maintaining processing simplicity through rule-based parameter selection.
Solution Approach 2:
The invention dynamically changes the weighting parameters according to the intra prediction type. By adjusting these parameters based on the prediction mode (planar, angular, DC), the system optimizes image quality for each local region without requiring complex adaptive algorithms, thus balancing simplicity and quality preservation.
Data Source
AI summary
More suitable image encoding technique and image decoding technique are provided. An image encoding method for encoding an image includes: a predicted image generation step of generating a predicted image of a synthesis prediction by performing synthesis processing of synthesizing a predicted image of an inter prediction and a predicted image of an intra prediction to an encoding-target block; and an encoding step of encoding a difference between the predicted image generated in the predicted image generation step and a pixel value of an image of the encoding-target block, the synthesizing processing includes weighting processing performed to the predicted image of the inter prediction and the predicted image of the intra prediction, and a weighting parameter of the weighting processing is determined in accordance with a type of the intra prediction.