-
[pdf]
[bibtex]@InProceedings{Conde_2023_CVPR, author = {Conde, Marcos V. and Zamfir, Eduard and Timofte, Radu and Motilla, Daniel and Liu, Cen and Zhang, Zexin and Peng, Yunbo and Lin, Yue and Guo, Jiaming and Zou, Xueyi and Chen, Yuyi and Liu, Yi and Hao, Jia and Yan, Youliang and Zhang, Yuanfan and Li, Gen and Sun, Lei and Kong, Lingshun and Bai, Haoran and Pan, Jinshan and Dong, Jiangxin and Tang, Jinhui and Ayazoglu, Mustafa and Bilecen, Bahri Batuhan and Li, Mingxi and Zhang, Yuhang and Fan, Xianjun and Sheng, Yankai and Sun, Long and Liu, Zibin and Gou, Weiran and Li, Shaoqing and Yi, Ziyao and Xiang, Yan and Kong, Dehui and Xu, Ke and Gankhuyag, Ganzorig and Yoon, Kihwan and Zhang, Jin and Yu, Gaocheng and Zhang, Feng and Wang, Hongbin and Zhou, Zhou and Chao, Jiahao and Gao, Hongfan and Gong, Jiali and Yang, Zhengfeng and Zeng, Zhenbing and Chen, Chengpeng and Guo, Zichao and Park, Anjin and Liu, Yuqing and Jia, Qi and Yu, Hongyuan and Yin, Xuanwu and Zuo, Kunlong and Zhang, Dongyang and Fu, Ting and Cheng, Zhengxue and Zhu, Shiai and Zhou, Dajiang and Yu, Weichen and Ge, Lin and Dong, Jiahua and Zou, Yajun and Wu, Zhuoyuan and Han, Binnan and Zhang, Xiaolin and Zhang, Heng and Shao, Ben and Zheng, Shaolong and Yin, Daheng and Chen, Baijun and Liu, Mengyang and Nistor, Marian-Sergiu and Chen, Yi-Chung and Huang, Zhi-Kai and Chiang, Yuan-Chun and Chen, Wei-Ting and Yang, Hao-Hsiang and Chang, Hua-En and Chen, I-Hsiang and Hsieh, Chia-Hsuan and Kuo, Sy-Yen and Vo, Tu and Yan, Qingsen and Zhu, Yun and Su, Jinqiu and Zhang, Yanning and Zhang, Cheng and Luo, Jiaying and Cho, Youngsun and Lee, Nakyung}, title = {Efficient Deep Models for Real-Time 4K Image Super-Resolution. NTIRE 2023 Benchmark and Report}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2023}, pages = {1495-1521} }
Efficient Deep Models for Real-Time 4K Image Super-Resolution. NTIRE 2023 Benchmark and Report
Abstract
This paper introduces a novel benchmark for efficient upscaling as part of the NTIRE 2023 Real-Time Image Super-Resolution (RTSR) Challenge, which aimed to upscale images from 720p and 1080p resolution to native 4K (x2 and x3 factors) in real-time on commercial GPUs. For this, we use a new test set containing diverse 4K images ranging from digital art to gaming and photography. We assessed the methods devised for 4K SR by measuring their runtime, parameters, and FLOPs, while ensuring a minimum PSNR fidelity over Bicubic interpolation. Out of the 170 participants, 25 teams contributed to this report, making it the most comprehensive benchmark to date and showcasing the latest advancements in real-time SR.
Related Material