Single Image Reflection Separation on SIR2 Objects (test)
27.11PSNRRDNet
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RDNetVenue=CVPR 2025, Trainable Parameters=266.4M2026.01 | 27.11 | 92.5 | |
| ReflexSplitVenue=–, Trainable Parameters=174M2026.01 | 27.08 | 92.9 | |
| DSITVenue=NeurIPS 2024, Trainable Parameters=136M2026.01 | 26.77 | 92.1 | |
| DSRNetVenue=ICCV 2023, Trainable Parameters=123.6M2026.01 | 26.74 | 92.1 | |
| Zhong et al.Venue=CVPR 2024, Trainable Parameters=–2026.01 | 26.51 | 92.7 | |
| MaxRFVenue=CVPR 2024, Trainable Parameters=27.9M2026.01 | 25.8 | 91.5 | |
| DExNetVenue=TPAMI 2025, Trainable Parameters=9.6M2026.01 | 25.37 | 90.9 | |
| YTMTVenue=NeurIPS 2021, Trainable Parameters=29.9M2026.01 | 25.03 | 89.8 | |
| Zhang et al.Venue=CVPR 2018, Trainable Parameters=3.1M2026.01 | 24.27 | 88.9 | |
| ERRNetVenue=CVPR 2019, Trainable Parameters=28.8M2026.01 | 24.1 | 89.1 | |
| IBCLNVenue=CVPR 2020, Trainable Parameters=24.3M2026.01 | 24.01 | 88.4 | |
| Zheng et al.Venue=CVPR 2021, Trainable Parameters=38.4M2026.01 | 22.48 | 81.3 |