Asymmetric Blending Masks for Heterogeneous Image Partitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding technologies, such as AV1, face challenges with interpolation that result in suboptimal blending of image partitions, particularly when dealing with heterogeneous content types.
Innovation Solution
Implement asymmetric blending masks with different thresholds for different image partitions, allowing for tailored blending based on the content type, enhancing geometric partitioning modes in codecs like VVC and AV2.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If symmetric blending masks with uniform thresholds are used for image partitioning, then the encoding process is simple and fast, but the blending accuracy deteriorates when dealing with heterogeneous content types
Solution Approach 1:
The patent applies different blending masks with different thresholds to different partitions of the same image block. Specifically, a first blending mask with a first threshold is applied to a first partition, while a second blending mask with a second threshold is applied to a second partition. This allows each partition to be blended with quality optimized for its specific content characteristics, thereby improving overall blending accuracy without requiring complex global processing
Solution Approach 2:
The patent divides the image block into multiple partitions and applies different blending strategies to each partition. By segmenting the blending process into partition-specific operations with customized thresholds, the system achieves high blending accuracy for heterogeneous content while maintaining relatively simple processing for each individual partition
2Measurement precision
If asymmetric blending masks with different thresholds are applied to different partitions, then blending accuracy for heterogeneous content is improved, but computational complexity increases
Solution Approach 1:
Instead of applying complex asymmetric blending to the entire image block, the patent applies simplified blending operations with partition-specific thresholds only to individual partitions. This localized approach improves blending accuracy for heterogeneous content while keeping the computational power required for each partition relatively low
3Measurement precision
If uniform blending thresholds are used across all partitions, then the encoding and decoding process is efficient, but the quality of reconstructed images deteriorates for heterogeneous content
Solution Approach 1:
The patent enhances image reconstruction quality by applying different blending thresholds to different partitions based on their content characteristics. Each partition undergoes blending processing optimized for its specific content type, improving overall reconstruction quality while maintaining encoding efficiency through the use of relatively simple threshold-based operations for each partition
4Adaptability or versatility
If partition-specific blending masks are implemented, then adaptability to different content types is improved, but the device complexity increases
Solution Approach 1:
The patent achieves content type adaptability by implementing partition-specific blending masks with different thresholds for different partitions. This allows the system to adapt to heterogeneous content types within a single image block while keeping the blending processing complexity relatively low by using simple threshold-based operations for each partition
Solution Approach 2:
By segmenting the image block into multiple partitions and applying different blending strategies to each, the system achieves high adaptability to different content types. The segmentation approach allows independent optimization for each partition without requiring complex global processing, thereby managing device complexity effectively
Data Source
AI summary
A decoding device is provided for decoding of a bitstream, and more specifically for predicting a picture area of an input image for decoding. The picture area has been divided into at least first and second parts by a partitioning boundary, and the pixels of each part are predicted according to suitable measures. Blending masks are then applied to generate blended regions for the predicted pixels, modifying the pixels to generate a complete prediction of the picture area based on both parts. The blending mask is based on first and second thresholds, which are defined relative to the partitioning boundary, and which may have different values to produce an asymmetrical blending relative to the boundary. Based on suitable threshold values, the prediction, combination, and decoding of the picture area is more adaptive to differing contents of the parts therein.