Neural Network Image Correction for Display Panel Seam Elimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large display devices face issues with display unevenness due to variations in transistor characteristics and display element characteristics, leading to noticeable boundaries between panels, especially in high-resolution displays with multiple panels.
Innovation Solution
A machine learning method using a neural network is employed to correct image data input to the display device, where the neural network learns to compensate for these variations by generating and updating weight coefficients based on differences between input and displayed images, ensuring seamless image continuity across panels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the display region is increased to provide larger screens, then the screen size and information display capacity are improved, but display unevenness and noticeable boundaries between panels occur due to transistor and display element characteristic variations
Solution Approach 1:
The patent applies local quality by dividing the display region into multiple panels with individually optimized characteristics. Each panel is treated as a separate unit with its own correction parameters, allowing localized compensation for transistor and display element variations. This enables the large display region to maintain uniformity across different panels through panel-specific correction data stored in memory.
2Area of stationary object
If multiple display panels are arranged to increase display region, then screen size is improved, but noticeable boundaries and seams appear between panels due to characteristic variations
Solution Approach 1:
The patent introduces correction data as an intermediary element between the multiple display panels. This correction data, stored in memory, acts as a mediator that compensates for the boundaries and seams between panels. By applying panel-specific correction parameters to the image signals before display, the system eliminates visible boundaries while maintaining the multi-panel configuration for larger display regions.
3Measurement precision
If high-resolution display is implemented, then image quality is improved, but transistor and display element characteristic variations become more noticeable
Solution Approach 1:
The patent applies preliminary action by pre-measuring and storing correction data for each display panel before actual operation. The characteristic variations of transistors and display elements are measured in advance, and correction parameters are calculated and stored in memory. During high-resolution display operation, these pre-prepared correction data are applied to compensate for the variations, allowing the system to maintain both high resolution and uniformity without real-time measurement overhead.
Data Source
AI summary
To improve the display quality of a display device. To provide a method of correcting image data input to the display device. To provide a novel image correction method or an image correction system. Machine learning for a neural network correcting image data input to the display device is performed by the following method: second image data based on an image that is displayed on the display device by input of first image data to the display device is obtained; third image data is generated by obtaining a difference between the first image data and the second image data; fourth image data is generated by adding the first image data and the third image data; and a weight coefficient is updated so that output data obtained by input of the first image data to the neural network is close to the fourth image data.