Encoder-Decoder Image Reconstruction for 4K/8K Edge Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image super-resolution techniques are inadequate for reconstructing images with 4K or 8K resolution and fail to preserve finer details such as image edges and textures, leading to insufficient image quality and increased storage and device costs.
Innovation Solution
A method utilizing a GAN-like variational autoencoder architecture (Soft-Intro VAE) that extracts features from low-resolution and original images, incorporating reference image details for high-definition reconstruction, enabling generation of high-resolution images with improved edge and texture preservation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If current super-resolution techniques are used for low-resolution upsampling, then processing speed is maintained, but image quality and detail preservation are insufficient for 4K or 8K resolution reconstruction
Solution Approach 1:
The patent segments the image processing task into multiple stages: extracting features from low-resolution images, obtaining reference features from high-resolution images, fusing these features, and reconstructing the super-resolution image. This segmentation allows each stage to be optimized independently, achieving both high quality and efficiency
Solution Approach 2:
The patent performs preliminary feature extraction from both low-resolution and high-resolution reference images before the actual reconstruction process. By pre-extracting and storing essential features, the system reduces computational burden during real-time processing while maintaining reconstruction quality
2Manufacturing precision
If high-resolution images are stored and processed directly, then image quality is preserved, but storage costs and device requirements increase
Solution Approach 1:
The patent extracts only the essential features from high-resolution reference images rather than storing and processing the complete high-resolution images. This extraction approach captures the critical information needed for reconstruction while dramatically reducing storage requirements
Solution Approach 2:
The patent creates a compressed feature representation (a simplified copy) of the high-resolution image information that can be used for reconstruction without requiring the original high-resolution data. This feature copy contains sufficient detail for quality reconstruction but occupies minimal storage space
3Quantity of substance
If low-resolution images are stored instead of high-resolution images, then storage costs are reduced, but image detail and edge preservation are lost
Solution Approach 1:
The patent introduces feature extraction and fusion as intermediary processes between the stored low-resolution images and the final high-resolution output. These intermediaries capture and preserve critical detail information that would otherwise be lost, enabling reconstruction of fine edges and textures without storing the original high-resolution data
Data Source
AI summary
A method in one embodiment includes: extracting first image features of a first image of a first resolution and second image features of a second image of a second resolution, wherein the first resolution is less than the second resolution, and the first image and the second image correspond to each other. The method further includes: extracting reference image features of a reference image, wherein the reference image includes edges and modes for reconstruction reference. The method further includes: generating a third image of a third resolution based on the first image features, the second image features, and the reference image features, wherein the third resolution is less than or equal to the second resolution. By using this method, super-resolution image processing can be implemented on a low-resolution image to obtain a reconstructed image with similar or the same resolution as the original image, with reduced storage space and device costs.