Occluded Object Recognition on Separated COCO (test)
31.15Top-1 Accuracypix2gestalt + CLIP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| pix2gestalt + CLIPEvaluation Mode=Zero-shot2024.01 | 31.15 | 45.77 | |
| CLIPEvaluation Mode=Zero-shot2024.01 | 26.04 | 43.19 | |
| CLIP + RCEvaluation Mode=Zero-shot2024.01 | 25.64 | 43.24 | |
| Vis. Obj. + CLIPEvaluation Mode=Zero-shot2024.01 | 21.1 | 34.7 |