FBRNN: Feedback Recurrent Neural Network for Extreme Image Super-Resolution
Abstract
Single image extreme Super Resolution (SR) is a difficult task as scale factor in the order of 10X or greater is typically attempted. For instance, in the case of 16x upscale of an image, a single pixel from a low resolution image gets expanded to a 16x16 image patch. Such attempts often result fuzzy quality and loss in details in reconstructed images. To handle these difficulties, we propose a network architecture composed of a series of connected blocks in recurrent and feedback fashions for enhanced SR reconstruction. By use of recurrent network, an SR image is refined over a sequence of enhancement stages in coarse to fine manner. Additionally, each stage involves back projection of SR image to LR images for continuously being refined during the sequence. According to the preliminary results of NTIRE 2020 Perceptual Extreme SR challenge, our team (KU_ISPLB) secured 6th place by PSNR and 7th place by SSIM among all participants.
Cite
Text
Lee et al. "FBRNN: Feedback Recurrent Neural Network for Extreme Image Super-Resolution." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020. doi:10.1109/CVPRW50498.2020.00252Markdown
[Lee et al. "FBRNN: Feedback Recurrent Neural Network for Extreme Image Super-Resolution." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020.](https://mlanthology.org/cvprw/2020/lee2020cvprw-fbrnn/) doi:10.1109/CVPRW50498.2020.00252BibTeX
@inproceedings{lee2020cvprw-fbrnn,
title = {{FBRNN: Feedback Recurrent Neural Network for Extreme Image Super-Resolution}},
author = {Lee, Junyeop and Park, Jaihyun and Lee, Kanghyu and Min, Jeongki and Kim, Gwantae and Lee, Bokyeung and Ku, Bonhwa and Han, David K. and Ko, Hanseok},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
year = {2020},
pages = {2021-2028},
doi = {10.1109/CVPRW50498.2020.00252},
url = {https://mlanthology.org/cvprw/2020/lee2020cvprw-fbrnn/}
}