Under-display Camera Image Preprocessing for Restoration Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image preprocessing methods for under-display cameras struggle to provide high-quality image data for effective image restoration training, due to the challenges of capturing images with deteriorated quality caused by the display pattern.
Innovation Solution
An image preprocessing device and method that generate an image data pair for image restoration training by receiving first and second image data, generating reference image data and crop image data, calculating loss function values, and selecting crop image data based on the smallest loss function value to output as the image data pair.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If an under-display camera is used to maintain display visibility, then the display area can be fully utilized, but image quality deteriorates due to the display pattern
Solution Approach 1:
The system performs preliminary actions by capturing multiple images at different positions and performing preprocessing (cropping, resizing, noise removal) before restoration training. This prepares the data in advance to compensate for the inherent quality deterioration caused by the under-display camera configuration.
Solution Approach 2:
The system creates multiple copies of image data by capturing images at different positions and generating crop images from these captures. These copied and processed images are then used for restoration training to compensate for the original image quality degradation.
2Manufacturing precision
If multiple image captures are performed for preprocessing, then image restoration training quality improves, but processing time increases
Solution Approach 1:
The system segments the image processing task by dividing it into distinct stages: capturing multiple images at different positions, preprocessing each capture (cropping, resizing, noise removal), calculating loss function values, and selecting optimal images for restoration training. This segmentation allows for efficient processing of multiple captures.
Solution Approach 2:
The system changes parameters by capturing images at different positions and adjusting preprocessing parameters (crop regions, resize dimensions, noise removal thresholds) to optimize the balance between training quality and processing efficiency. Loss function values are calculated to guide parameter selection.
Data Source
AI summary
An image data pair for image restoration training is generated by an image preprocessing method. The image preprocessing method includes receiving first image data and second image data, generating reference image data based on the first image data, generating a plurality of crop image data based on the second image data, selecting crop image data based on a smallest loss function value among loss function values generated based on each of the reference image data and the plurality of crop image data, and outputting the reference image data and the selected crop image data as the image data pair.