Deep RAW 图像超分辨率：NTIRE 2024 挑战综述
计算机视觉与模式识别
2024-04-26 v1 图像与视频处理
摘要
本文综述了 NTIRE 2024 RAW 图像超分辨率挑战，概述了所提出的方法和结果。对于现代图像信号处理 (ISP) 管道而言，新的 RAW 超分辨率方法可能至关重要，但该问题在 RGB 领域之外并未得到充分探索。本次挑战旨在将 RAW Bayer 图像放大 2 倍，考虑未知退化如噪声和模糊。在本次挑战中，共有 230 名参赛者报名参赛，期间收到了 45 份提交结果。本文审查了前 5 名提交结果的性能，作为当前 RAW 图像超分辨率最先进水平的衡量标准。
引用
@article{arxiv.2404.16223,
title = {Deep RAW Image Super-Resolution. A NTIRE 2024 Challenge Survey},
author = {Marcos V. Conde and Florin-Alexandru Vasluianu and Radu Timofte and Jianxing Zhang and Jia Li and Fan Wang and Xiaopeng Li and Zikun Liu and Hyunhee Park and Sejun Song and Changho Kim and Zhijuan Huang and Hongyuan Yu and Cheng Wan and Wending Xiang and Jiamin Lin and Hang Zhong and Qiaosong Zhang and Yue Sun and Xuanwu Yin and Kunlong Zuo and Senyan Xu and Siyuan Jiang and Zhijing Sun and Jiaying Zhu and Liangyan Li and Ke Chen and Yunzhe Li and Yimo Ning and Guanhua Zhao and Jun Chen and Jinyang Yu and Kele Xu and Qisheng Xu and Yong Dou},
journal= {arXiv preprint arXiv:2404.16223},
year = {2024}
}
备注
CVPR 2024 - NTIRE Workshop