-
[pdf]
[arXiv]
[bibtex]@InProceedings{Conde_2024_CVPR, author = {Conde, Marcos V. and Lei, Zhijun and Li, Wen and Katsavounidis, Ioannis and Timofte, Radu and Yan, Min and Liu, Xin and Wang, Qian and Ye, Xiaoqian and Du, Zhan and Zhang, Tiansen and Li, Zhiyuan and Wei, Hao and Ge, Chenyang and Lv, Jiangtao and Sun, Long and Pan, Jinshan and Dong, Jiangxin and Tang, Jinhui and Zhou, Menghan and Yan, Yiqiang and Yoon, Kihwan and Gankhuyag, Ganzorig and Lee, Jae-Hyeon and Choi, Ui-Jin and Moon, Hyeon-Cheol and Jeong, Tae-Hyun and Yang, Yoonmo and Kim, Jae-Gon and Jeong, Jinwoo and Kim, Sunjei and Qiu, Xintao and Zhou, Yuanbo and Wu, Kongxian and Dai, Xinwei and Tang, Hui and Deng, Wei and Gao, Qingquan and Tong, Tong and Peng, Long and Guo, Jiaming and Di, Xin and Liao, Bohao and Du, Zhibo and Xia, Peize and Pei, Renjing and Wang, Yang and Cao, Yang and Zha, Zhengjun and Han, Bingnan and Yu, Hongyuan and Wu, Zhuoyuan and Wan, Cheng and Liu, Yuqing and Yu, Haodong and Li, Jizhe and Huang, Zhijuan and Huang, Yuan and Zou, Yajun and Guan, Xianyu and Jia, Qi and Zhang, Heng and Yin, Xuanwu and Zuo, Kunlong and Zhang, Dongyang and Liu, Tianle and Chen, Huaian and Jin, Yi}, title = {Real-Time 4K Super-Resolution of Compressed AVIF Images. AIS 2024 Challenge Survey}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2024}, pages = {5838-5856} }
Real-Time 4K Super-Resolution of Compressed AVIF Images. AIS 2024 Challenge Survey
Abstract
This paper introduces a novel benchmark for efficient image upscaling as part of the AIS 2024 Real-Time Image Super-Resolution (RTSR) Challenge which aims to upscale compressed images from 540p to native 4K resolution (4x factor) in real-time on commercial GPUs. For this we use a diverse test set containing diverse 4K images ranging from digital art to gaming and photography. The images are compressed using the modern AVIF codec instead of JPEG. All the proposed methods improve PSNR fidelity over Lanczos interpolation and process images under 30ms. Out of the 160 participants 25 teams submitted their code. This survey considers only the most novel solutions models making it the most comprehensive benchmark on real-time SR of compressed images using modern codecs.
Related Material