28‐3: Frequency Decomposition‐Based High‐Performance Demura Processing with Low Memory Cost

In this paper, we propose a new demura method for constructing display mura profiles by decomposing sampled mura data from the entire gray level into representative high spatial frequency components and multiple low spatial frequency components. By effectively allocating the components during demura processing, we achieved 98.5% in luminance uniformity while maintaining the color accuracy. This approach allows for implementation with 65.6% reduced memory size for compensation data storage compared to the conventional method, making it more cost‐efficient in the display driver IC (DDIC).