Multi-Channel Convolutional Neural Networks for Image Super-Resolution

Special Section on Image Media Quality

Shinya OHTANI, Yu KATO, Nobutaka KUROKI, Tetsuya HIROSE, Masahiro NUMA

Author information

Keywords: super-resolution, resolution enhancement, convolutional neural networks, CNN, deep learning

JOURNAL RESTRICTED ACCESS

2017 Volume E100.A Issue 2 Pages 572-580

DOI https://doi.org/10.1587/transfun.E100.A.572

Details

Abstract

This paper proposes image super-resolution techniques with multi-channel convolutional neural networks. In the proposed method, output pixels are classified into K×K groups depending on their coordinates. Those groups are generated from separate channels of a convolutional neural network (CNN). Finally, they are synthesized into a K×K magnified image. This architecture can enlarge images directly without bicubic interpolation. Experimental results of 2×2, 3×3, and 4×4 magnifications have shown that the average PSNR for the proposed method is about 0.2dB higher than that for the conventional SRCNN.

Corresponding author

Register with J-STAGE for free!