International Display Workshops General Incorporated Association

13:30 〜 13:50

[DES4-1L] CNN-based Subpixel Rendering Technique with Additional Pixel Arrangement Information

Yun Jang Pyun1, *Garam Kim1, Hyoungsik Nam1 (1. Kyung Hee University (Korea))

subpixel rendering, deep neural network, apparent resolution

https://doi.org/10.36463/idw.2023.1078

We present a subpixel rendering technique based on convolutional neural network (CNN), using additional pixel arrangement information. The proposed subpixel rendering network, SPRNN, provides apparently high resolution down-sampled image from the input image and pixel arrangement information. Experimental results show that the proposed SPRNN outperforms existing techniques.