Iceclear/StableSR-TestSets
收藏资源简介:
StableSR TestSets数据集由Jianyi Wang开发,包含合成和真实世界的测试集,用于图像超分辨率任务。该数据集的主要用途是复现论文中展示的度量结果。数据集包含多个测试集,如DIV2K_Val、RealSR Val、DRealSR Val和DPED Val,分别用于验证不同的图像超分辨率模型。数据集的许可证为S-Lab License 1.0。
The StableSR TestSets dataset was developed by Jianyi Wang. It comprises both synthetic and real-world test sets for image super-resolution tasks. The primary purpose of this dataset is to reproduce the metric results presented in the corresponding paper. The dataset includes multiple test subsets such as DIV2K_Val, RealSR Val, DRealSR Val and DPED Val, which are respectively used to validate different image super-resolution models. The dataset is licensed under S-Lab License 1.0.
StableSR 测试集卡片
数据详情
-
开发者: Jianyi Wang
-
数据类型: 合成和真实世界的图像超分辨率测试集
-
许可证: S-Lab License 1.0
-
数据描述: 这些测试集用于重现论文中展示的指标结果。
-
更多信息资源: GitHub 仓库
-
引用方式:
@InProceedings{wang2023exploiting, author = {Wang, Jianyi and Yue, Zongsheng and Zhou, Shangchen and Chan, Kelvin CK and Loy, Chen Change}, title = {Exploiting Diffusion Prior for Real-World Image Super-Resolution}, booktitle = {arXiv preprint arXiv:2305.07015}, year = {2023}, }
使用
目前提供的测试集包括:
- DIV2K_Val: 3000 对合成数据,用于DIV2K验证集,使用与训练 StableSR 相同的退化方法生成。
- RealSR Val: 中心裁剪的数据对,来自RealSRv3。
- DRealSR Val: 中心裁剪的数据对,来自DRealSR。
- DPED Val: 中心裁剪的低质量仅数据,来自DPED。
评估结果
详细评估结果请参见论文。




