Surgical Video Recognition on Cat-21
15.69Top-1 AccuracyDINOv3
Evaluation Results
| Method | Links | |
|---|---|---|
| DINOv3Pre-training Domain=Generic2026.03 | 15.69 | |
| DINOv3Pre-training Domain=Surgical2026.03 | 15.69 | |
| VideoMAEPre-training Domain=Generic2026.03 | 15.69 | |
| SurgRec-MAEBalanced Sampling=false2026.03 | 15.69 | |
| JEPAPre-training Domain=Generic2026.03 | 15.69 | |
| SurgRec-MAEBalanced Sampling=true2026.03 | 15.69 | |
| SurgRec-JEPABalanced Sampling=true2026.03 | 15.69 | |
| SurgRec-MAE2026.03 | 15.69 | |
| Qwen2.5-VL-7BPrompt strategy=P12026.03 | 13.73 | |
| Qwen2.5-VL-7BPrompt strategy=P22026.03 | 13.73 | |
| LLaVA-NeXT-7BPrompt strategy=P22026.03 | 11.76 | |
| LLaVA-NeXT-7BPrompt strategy=P12026.03 | 9.8 | |
| Qwen3-VL-8BPrompt strategy=P12026.03 | 7.84 | |
| Qwen3-VL-8BPrompt strategy=P22026.03 | 3.92 |