NTIRE 2021 Learning the Super-Resolution Space Challenge
NTIRE 2021 Learning the Super-Resolution Space Challenge
复制标题
DOI：
10.1109/cvprw53098.2021.00072
复制
发表时间：
2021-06
期刊：
影响因子：
--
通讯作者：
Andreas Lugmayr;Martin Danelljan;R. Timofte;C. Busch;Yang Chen;Jian Cheng;Vishal M. Chudasama;Ruipeng Gang;Shangqi Gao;Kun Gao;Laiyun Gong;Qingrui Han NetEaseYunXin;Chao Huang;Zhi Jin;Younghyun Jo;Seon Joo Kim;Younggeun Kim;Seungjun Lee;Yu Lei;Chu-Tak Li;Chenghua Li;Ke Li;Zhi-Song Liu;Youming Liu;Nan Nan-Nan;Seung-Ho Park;Heena Patel;Shichong Peng;Kalpesh P. Prajapati;Haoran Qi;K. Raja;Raghavendra Ramachandra;W. Siu;Donghee Son;Ruixia Song;K. Upla;Li-Wen Wang;Yatian Wang;Junwei Wang;Qianyu Wu;Xinhua Xu;Sejong Yang;Zhen Yuan NetEaseYunXin;Liting Zhang;Huanrong Zhang;Junkai Zhang;Yifan Zhang;Zhenzhou Zhang;Hang Zhou;A. Zhu;X. Zhuang;Jiaxin Zou
中科院分区：
文献类型：
--
作者：
Andreas Lugmayr;Martin Danelljan;R. Timofte;C. Busch;Yang Chen;Jian Cheng;Vishal M. Chudasama;Ruipeng Gang;Shangqi Gao;Kun Gao;Laiyun Gong;Qingrui Han NetEaseYunXin;Chao Huang;Zhi Jin;Younghyun Jo;Seon Joo Kim;Younggeun Kim;Seungjun Lee;Yu Lei;Chu-Tak Li;Chenghua Li;Ke Li;Zhi-Song Liu;Youming Liu;Nan Nan-Nan;Seung-Ho Park;Heena Patel;Shichong Peng;Kalpesh P. Prajapati;Haoran Qi;K. Raja;Raghavendra Ramachandra;W. Siu;Donghee Son;Ruixia Song;K. Upla;Li-Wen Wang;Yatian Wang;Junwei Wang;Qianyu Wu;Xinhua Xu;Sejong Yang;Zhen Yuan NetEaseYunXin;Liting Zhang;Huanrong Zhang;Junkai Zhang;Yifan Zhang;Zhenzhou Zhang;Hang Zhou;A. Zhu;X. Zhuang;Jiaxin Zou
本文回顾了2021年NTIRE关于学习超分辨率空间的挑战赛。它聚焦于参赛方法和最终结果。该挑战赛致力于解决从单张低分辨率图像中学习一个能够预测合理超分辨率（SR）图像空间的模型这一问题。因此，该模型必须能够对不同的输出进行采样，而不仅仅是生成单张超分辨率图像。挑战赛的目标是激励研究人员开发更适合于高度不适定超分辨率问题的学习公式和模型，从而推动更广泛的超分辨率领域的技术发展水平。为了评估预测的超分辨率空间的质量，我们提出了一种新的评估指标，并对参赛方法进行了全面分析。挑战赛包含两个赛道：4倍和8倍缩放因子。共有11支队伍参加了最终的测试阶段。
This paper reviews the NTIRE 2021 challenge on learning the super-Resolution space. It focuses on the participating methods and final results. The challenge addresses the problem of learning a model capable of predicting the space of plausible super-resolution (SR) images, from a single low-resolution image. The model must thus be capable of sampling diverse outputs, rather than just generating a single SR image. The goal of the challenge is to spur research into developing learning formulations and models better suited for the highly ill-posed SR problem. And thereby advance the state-of-the-art in the broader SR field. In order to evaluate the quality of the predicted SR space, we propose a new evaluation metric and perform a comprehensive analysis of the participating methods. The challenge contains two tracks: 4× and 8 scale factor. In total, 11 teams competed in the final testing× phase.