For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Multi-Channel Convolutional Neural Networks for Image Super-Resolution
Shinya OHTANI Yu KATO Nobutaka KUROKI Tetsuya HIROSE Masahiro NUMA
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 2017/02/01
Online ISSN: 1745-1337
Type of Manuscript: Special Section PAPER (Special Section on Image Media Quality)
Category: IMAGE PROCESSING
super-resolution, resolution enhancement, convolutional neural networks, CNN, deep learning,
Full Text: PDF>>
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.