Video Segmentation on 140 Gestalt videos
94AccuracyGenMatter
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GenMatterResolution=96x96, Evaluation protocol=probe-point sampling, Inputs=optical flow and depth, Inference type=probabilistic inference2026.04 | 94 | 72 | |
| FlowSAMResolution=96x96, Evaluation protocol=probe-point sampling, Inputs=optical flow2026.04 | 87 | 67 | |
| SegAnyMoResolution=96x96, Evaluation protocol=probe-point sampling, Inputs=point tracking and monocular depth2026.04 | 33 | 26 |