Image Decompression Fusion of Y and UV Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image/video compression methods using convolutional neural networks face inefficiencies due to separate processing paths for Y and UV components, leading to high computational complexity and reduced performance.
Innovation Solution
An image decompression method that optimizes the fusion of Y and UV components by performing concatenation and convolution operations, allowing for flexible channel management to reduce computational load while enhancing performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If separate processing paths are used for Y and UV components, then processing simplicity is maintained, but computational complexity increases and performance decreases
Solution Approach 1:
The patent merges the separate processing paths for Y and UV components into a unified processing framework. The Y component and UV component are processed through shared convolutional layers and feature fusion mechanisms, reducing overall computational complexity while maintaining processing simplicity. This integration allows the system to achieve better compression performance without increasing device complexity.
2Power
If channel quantity is reduced to save computing power, then computational load decreases, but performance may deteriorate
Solution Approach 1:
The patent dynamically adjusts the channel quantity parameter in different processing stages. In early stages, a larger channel quantity is used to capture detailed features. In later stages, the channel quantity is reduced through feature fusion and refinement operations, achieving computational efficiency without sacrificing performance. The specific channel dimensions are optimized based on the processing requirements at each stage.
Solution Approach 2:
The patent applies partial channel reduction strategy where not all channels are uniformly reduced. Instead, selective channel pruning is performed based on feature importance, maintaining essential channels while removing redundant ones. This partial action approach preserves compression performance while significantly reducing computational power requirements.
3Device complexity
If Y and UV components are processed separately, then processing simplicity is maintained, but information exchange between components is insufficient
Solution Approach 1:
The patent introduces feature fusion modules as intermediaries between the Y component processing path and UV component processing path. These fusion modules enable bidirectional information exchange by combining features from both components and feeding them back to respective processing paths. This intermediary mechanism ensures sufficient information exchange while maintaining overall processing simplicity through modular architecture.
Data Source
AI summary
This application provides an image decompression method and apparatus. The image decompression method in this application includes: obtaining a first feature tensor, where the first feature tensor corresponds to a first component of a reconstructed image; obtaining a second feature tensor, where the second feature tensor corresponds to a second component of the reconstructed image; performing concatenation and convolution on the first feature tensor and the second feature tensor to obtain a third feature tensor; obtaining a fourth feature tensor based on the third feature tensor and the second feature tensor; and obtaining the reconstructed image based on the first feature tensor and the fourth feature tensor. In embodiments of this application, a structure for fusing a Y component and a UV component is optimized.