Super-Resolution via Conditional Implicit Maximum Likelihood Estimation
Machine Learning
2018-10-03 v1 Computer Vision and Pattern Recognition
Graphics
Neural and Evolutionary Computing
Machine Learning
Abstract
Single-image super-resolution (SISR) is a canonical problem with diverse applications. Leading methods like SRGAN produce images that contain various artifacts, such as high-frequency noise, hallucinated colours and shape distortions, which adversely affect the realism of the result. In this paper, we propose an alternative approach based on an extension of the method of Implicit Maximum Likelihood Estimation (IMLE). We demonstrate greater effectiveness at noise reduction and preservation of the original colours and shapes, yielding more realistic super-resolved images.
Cite
@article{arxiv.1810.01406,
title = {Super-Resolution via Conditional Implicit Maximum Likelihood Estimation},
author = {Ke Li and Shichong Peng and Jitendra Malik},
journal= {arXiv preprint arXiv:1810.01406},
year = {2018}
}
Comments
12 pages, 7 figures