Image Matting On Aim 500
Metrics
Conn.
Grad.
MAD
MSE
SAD
Results
Performance results of various models on this benchmark
| Paper Title | ||||||
|---|---|---|---|---|---|---|
| LF | 181.26 | 63.51 | 0.1130 | 0.0667 | 191.74 | A Late Fusion CNN for Digital Matting |
| SHM | 170.67 | 115.29 | 0.1012 | 0.0921 | 170.44 | Semantic Human Matting |
| U2NET | 82.14 | 51.02 | 0.0493 | 0.0348 | 83.46 | U2-Net: A Bayesian U-Net model with epistemic uncertainty feedback for photoreceptor layer segmentation in pathological OCT scans |
| GFM | 52.69 | 46.11 | 0.0313 | 0.0213 | 52.66 | Bridging Composite and Real: Towards End-to-end Deep Image Matting |
| AIM-Net | 43.18 | 33.05 | 0.0262 | 0.0161 | 43.92 | Deep Automatic Natural Image Matting |
| DiffMatte | 15.98 | 15.68 | 0.0098 | 0.0033 | 16.31 | Diffusion for Natural Image Matting |
0 of 6 row(s) selected.