Blended Neural Network Super-Resolution Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image data processing technologies face challenges in efficiently converting lower resolution images to higher resolution without introducing jagged artifacts, and methods to remove these artifacts often increase processing time.
Innovation Solution
An electronic device with an enhancement processor, neural network, feature detection processor, and blending logic circuit combines non-neural network and neural network enhanced image data to generate higher resolution output, using directional scaling and blending based on texture statistics to overcome the limitations of neural networks in super-resolution enhancement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If non-neural network image processing schemes are used to convert lower resolution images to higher resolution, then processing time is reduced, but jagged artifacts and errors appear in the enhanced image
Solution Approach 1:
The patent combines neural network processing and non-neural network processing into a unified super-resolution system. The neural network processes certain image regions to generate high-quality enhanced pixels, while non-neural network methods handle other regions, and the results are blended together to achieve both speed and quality.
Solution Approach 2:
The patent applies different processing methods to different regions of the image based on local characteristics. Regions with important features or higher quality requirements are processed using neural networks, while other regions use non-neural network methods, optimizing both quality and processing efficiency.
2Manufacturing precision
If neural network processing is used to enhance image resolution, then image quality improves, but processing time increases
Solution Approach 1:
The patent applies neural network processing selectively to only certain portions of the image rather than the entire image. This partial application of neural network processing maintains image quality in critical regions while reducing overall processing time through the blending with non-neural network processed regions.
3Manufacturing precision
If artifact removal operations are performed on enhanced images, then image quality improves, but processing time increases
Solution Approach 1:
The patent performs artifact reduction during the initial super-resolution enhancement process itself, rather than as a separate post-processing step. The neural network and blending operations inherently reduce artifacts while enhancing resolution, eliminating the need for additional artifact removal processing time.
Data Source
AI summary
Embodiments relate to a super-resolution engine that converts a lower resolution input image into a higher resolution output image. The super-resolution engine includes a directional scaler, an enhancement processor, a feature detection processor, a blending logic circuit, and a neural network. The directional scaler generates directionally scaled image data by upscaling the input image. The enhancement processor generates enhanced image data by applying an example-based enhancement, a peaking filter, or some other type of non-neural network image processing scheme to the directionally scaled image data. The feature detection processor determines features indicating properties of portions of the directionally scaled image data. The neural network generates residual values defining differences between a target result of the super-resolution enhancement and the directionally scaled image data. The blending logic circuit blends the enhanced image data with the residual values according to the features.