Referring Image Matting Refmatte Rw100 On
평가 지표
MAD
MAD(E)
MSE
MSE(E)
SAD
SAD(E)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | MAD | MAD(E) | MSE | MSE(E) | SAD | SAD(E) | Paper Title | Repository |
|---|---|---|---|---|---|---|---|---|
| MDETR (ResNet-101) | 0.0751 | 0.0779 | 0.0675 | 0.0700 | 131.58 | 136.59 | MDETR -- Modulated Detection for End-to-End Multi-Modal Understanding | |
| CLIPMat (ViT-L/14) | 0.0510 | 0.0505 | 0.0488 | 0.0483 | 88.52 | 87.92 | Referring Image Matting | |
| CLIPSeg (ViT-B/16) | 0.1222 | 0.1282 | 0.1178 | 0.1236 | 211.86 | 222.37 | Image Segmentation Using Text and Image Prompts | |
| CLIPMat (ViT-B/16) | 0.0636 | 0.0635 | 0.0614 | 0.0612 | 110.66 | 110.63 | Referring Image Matting |
0 of 4 row(s) selected.