AIM 2025挑战：逆 tone mapping 任务报告：方法与结果
计算机视觉与模式识别
2025-09-23 v2 图像与视频处理
摘要
本文综合回顾了AIM 2025挑战中关于逆 tone mapping（ITM）的工作。该挑战旨在推动开发有效的ITM算法,以从单一 LDR 输入重建 HDR 图像,关注感知保真度和数值一致性。共有67个参赛者提交了319个有效结果,其中最佳五支队伍被选为详细分析对象。本报告汇总了他们的方法和性能,其中最佳条目中最低的PU21-PSNR达到29.22 dB。分析结果表明,创新策略不断增强HDR重建质量,并为未来在逆 tone mapping领域的研究设定了强有力的基准。
引用
@article{arxiv.2508.13479,
title = {AIM 2025 challenge on Inverse Tone Mapping Report: Methods and Results},
author = {Chao Wang and Francesco Banterle and Bin Ren and Radu Timofte and Xin Lu and Yufeng Peng and Chengjie Ge and Zhijing Sun and Ziang Zhou and Zihao Li and Zishun Liao and Qiyu Kang and Xueyang Fu and Zheng-Jun Zha and Zhijing Sun and Xingbo Wang and Kean Liu and Senyan Xu and Yang Qiu and Yifan Ding and Gabriel Eilertsen and Jonas Unger and Zihao Wang and Ke Wu and Jinshan Pan and Zhen Liu and Zhongyang Li and Shuaicheng Liu and S. M Nadim Uddin},
journal= {arXiv preprint arXiv:2508.13479},
year = {2025}
}