Human Instance Segmentation
Instance segmentation is the task of detecting and delineating each distinct object of interest appearing in an image.
Image Credit: Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers
Papers
No papers found.
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BBox-Mask-Pose 2x | AP | 32.4 | — | Unverified |
| 2 | Crowd-SAM (ViT-L) | AP | 31.4 | — | Unverified |
| 3 | HQNet (ResNet-50) | AP | 31.1 | — | Unverified |
| 4 | BlendMask + CIS | AP | 29.8 | — | Unverified |
| 5 | Mask2Former + Occlusion C&P | AP | 28.3 | — | Unverified |
| 6 | CondInst + CIS | AP | 28.1 | — | Unverified |
| 7 | Mask2Former | AP | 27.8 | — | Unverified |
| 8 | HCQNet | AP | 27.3 | — | Unverified |
| 9 | ExPoSeg | AP | 26.8 | — | Unverified |
| 10 | RTMDet-ins-l | AP | 26.5 | — | Unverified |