Research Area: Machine Learning
Recently, it has been shown that in super-resolution, there exists a tradeoff relationship between the quantitative and perceptual quality of super-resolved images, which correspond to the similarity to the ground-truth images and the naturalness, respectively. In this paper, we propose a novel super-resolution method that can improve the perceptual quality of the upscaled images while preserving the conventional quantitative performance. The proposed method employs a deep network for multi-pass upscaling in company with a discriminator network and two qualitative score predictor networks. Experimental results demonstrate that the proposed method achieves a good balance of the quantitative and perceptual quality, showing more satisfactory results than existing methods.
Keywords:
Author(s) Name: Jun-Ho Choi, Jun-Hyuk Kim, Manri Cheon, Jong-Seok Lee
Journal name: Neurocomputing
Conferrence name:
Publisher name: Elsevier
DOI: 10.1016/j.neucom.2019.06.103
Volume Information: Volume 398, 20 July 2020, Pages 347-359
