Display Image Correction Using ML for Uneven Luminance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Display devices with large substrates face issues such as display unevenness due to differences in light exposure between light exposure regions, leading to noticeable bright spots and reduced image quality, which are difficult to address during manufacturing or post-manufacturing deterioration.
Innovation Solution
An image processing system utilizing a machine learning model generated by a learning device that processes image data to minimize display unevenness and bright spots by adjusting pixel luminance, incorporating a display device, image capturing device, and a learning device to analyze and correct image data based on captured images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If divided light exposure is performed to accommodate large substrates, then the substrate size can be increased, but display unevenness occurs due to deviation of mask position from light exposure regions
Solution Approach 1:
The patent applies preliminary action by measuring the actual luminance of each pixel during or after the light exposure process and using this information to pre-calculate correction values. These correction values are stored and applied during image processing to compensate for the non-uniform light exposure caused by mask position deviations, thereby maintaining display uniformity across large substrates without reducing substrate size.
2Manufacturing precision
If image processing is performed to correct display unevenness, then image quality is improved, but processing time increases
Solution Approach 1:
The patent performs the luminance measurement and creation of correction lookup tables during the manufacturing process rather than during runtime image processing. By pre-calculating and storing correction values for each pixel position, the actual image processing during display only requires simple table lookup and application, significantly reducing processing time while maintaining display uniformity correction.
Solution Approach 2:
The patent creates a lookup table that copies and stores pre-measured luminance data and correction values for each pixel position. During image processing, instead of performing complex calculations for each pixel, the system simply retrieves pre-computed correction values from this lookup table, dramatically reducing processing time while maintaining correction accuracy.
3Manufacturing precision
If machine learning based image processing is used, then display unevenness is reduced, but computational complexity increases
Solution Approach 1:
The patent simplifies the machine learning approach by creating a lookup table that directly stores pre-measured luminance data and correction values for each pixel position. This replaces complex machine learning algorithms with a simple data structure that allows fast retrieval and application of corrections, reducing computational complexity while maintaining effective display uniformity correction.
Solution Approach 2:
The patent extracts the essential correction information from complex image processing and isolates it into a separate lookup table structure. This separates the correction data from the image processing workflow, allowing simple table lookup operations instead of complex computations, thereby reducing device complexity while maintaining display uniformity improvement.
Data Source
AI summary
An image processing system that can reduce display unevenness in an image displayed on a display device is provided. The image processing system includes a display device, an image capturing device, and a learning device. The learning device stores a table representing information on the correspondence between first image data and second image data that is generated by display of an image corresponding to the first image data on the display device and image capturing of the image by the image capturing device. The learning device generates teacher data in accordance with the table and generates a machine learning model with the use of the teacher data generated. Image processing using the machine learning model is performed on image data input to the display device, so that display unevenness in the image displayed on the display device can be reduced.