Models | Translation (Dolly/Pedestal/Truck) | Zooming | Rotation (Pan/Tilt/Roll) | Static | Avg | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
In | Out | Up | Down | Right | Left | In | Out | Right | Left | Up | Down | CW | CCW | |||
🏅Qwen-2.5VL-7B (Ours SFT) | 83.2 | 48.6 | 27.2 | 48.8 | 62.6 | 54.3 | 51.3 | 70.7 | 77.6 | 86.9 | 70.4 | 58.0 | 38.5 | 46.3 | 65.2 | 59.3 |
🥈MegaSAM | 73.8 | 43.9 | 24.2 | 29.1 | 45.3 | 44.2 | 11.1 | 10.2 | 79.5 | 82.2 | 73.8 | 65.3 | 71.5 | 75.8 | 22.0 | 50.1 |
🥉MASt3R | 47.5 | 21.1 | 23.5 | 40.2 | 38.7 | 38.1 | 42.2 | 46.6 | 66.6 | 58.0 | 63.2 | 40.3 | 50.4 | 53.5 | 15.7 | 43.1 |
CUT3R | 68.9 | 50.4 | 24.7 | 34.2 | 37.0 | 27.6 | 15.9 | 21.3 | 59.1 | 65.0 | 65.0 | 47.5 | 60.7 | 66.2 | 15.1 | 42.7 |
VGGSFM | 56.6 | 28.9 | 28.7 | 38.2 | 48.9 | 35.3 | 21.7 | 17.3 | 60.9 | 58.7 | 46.6 | 43.3 | 61.4 | 55.5 | 16.7 | 41.3 |
DUSt3R | 58.9 | 24.0 | 30.7 | 18.0 | 38.3 | 26.9 | 18.2 | 24.6 | 59.4 | 63.8 | 32.9 | 27.3 | 61.0 | 57.9 | 13.1 | 37.0 |
GPT-4o | 66.3 | 29.2 | 21.1 | 38.2 | 38.0 | 21.9 | 41.7 | 39.3 | 44.7 | 42.1 | 43.6 | 35.5 | 24.0 | 28.7 | 32.0 | 36.4 |
InternVL2.5-26B | 63.6 | 11.8 | 21.1 | 23.6 | 27.2 | 19.4 | 21.8 | 31.6 | 42.5 | 38.3 | 44.9 | 43.6 | 14.3 | 18.2 | 25.1 | 29.8 |
Qwen2.5-VL-7B | 63.0 | 14.1 | 20.1 | 22.3 | 28.5 | 27.7 | 23.2 | 27.2 | 36.5 | 44.6 | 38.4 | 25.7 | 26.0 | 25.5 | 20.2 | 29.5 |
Qwen2.5-VL-32B | 66.8 | 19.1 | 11.1 | 31.4 | 32.1 | 30.4 | 27.8 | 32.6 | 43.2 | 50.0 | 53.2 | 44.0 | 26.6 | 29.0 | 28.8 | 35.1 |
Qwen2.5-VL-72B | 67.2 | 19.1 | 12.8 | 26.5 | 33.3 | 26.1 | 27.5 | 41.2 | 50.6 | 46.8 | 53.4 | 31.0 | 33.3 | 30.9 | 29.1 | 35.3 |
InternVL2.5-8B | 67.9 | 12.9 | 28.1 | 25.9 | 23.4 | 23.2 | 18.6 | 32.1 | 37.4 | 30.9 | 37.6 | 36.9 | 11.5 | 25.3 | 23.4 | 29.5 |
COLMAP | 36.2 | 13.1 | 11.9 | 19.7 | 34.1 | 30.0 | 13.9 | 14.2 | 43.9 | 46.4 | 28.3 | 19.1 | 42.1 | 48.7 | 7.5 | 27.3 |
LLAVA-Video-7B | 54.7 | 15.2 | 16.5 | 19.3 | 27.1 | 23.6 | 16.2 | 16.9 | 33.6 | 36.8 | 26.9 | 37.2 | 16.1 | 21.7 | 22.1 | 25.6 |
LLAVA-OneVision-7B | 46.8 | 13.5 | 12.6 | 16.9 | 23.7 | 20.2 | 10.7 | 14.4 | 33.5 | 33.6 | 16.9 | 31.4 | 19.3 | 20.8 | 18.8 | 22.2 |
InternVideo2-Chat-8B | 69.9 | 18.5 | 19.3 | 17.6 | 17.9 | 23.4 | 12.2 | 10.4 | 22.6 | 22.7 | 17.2 | 22.8 | 19.6 | 16.4 | 20.2 | 22.0 |
Tarsier-Recap-7B | 59.7 | 15.1 | 25.7 | 23.7 | 28.8 | 21.5 | 14.4 | 15.0 | 22.8 | 27.3 | 24.6 | 21.6 | 15.2 | 18.7 | 30.7 | 21.0 |
mPLUG-Owl3-7B | 47.6 | 12.9 | 13.9 | 16.9 | 17.3 | 18.5 | 12.9 | 10.6 | 31.4 | 26.6 | 26.1 | 37.0 | 10.4 | 12.2 | 17.8 | 20.8 |
InternLMXComposer2.5-7B | 49.0 | 10.6 | 11.4 | 10.4 | 14.6 | 10.6 | 11.8 | 16.5 | 14.3 | 13.9 | 14.7 | 17.5 | 11.7 | 18.1 | 21.8 | 16.5 |