| object-localization-on-grit | Mask R-CNN | #1 | Localization (ablation): 44.7Localization (test): 45.1 |
| instance-segmentation-on-bdd100k-val | Mask R-CNN | #3 | AP: 20.5 |
| multi-human-parsing-on-mhp-v20 | Mask R-CNN | #5 | AP 0.5: 14.9 |
| multi-person-pose-estimation-on-ochuman | Mask R-CNN | #8 | AP50: 33.2AP75: 24.5Validation AP: 20.2 |
| multi-tissue-nucleus-segmentation-on-kumar | Mask R-CNN (e) | #10 | Dice: 0.760Hausdorff Distance (mm): 50.9 |
| multi-person-pose-estimation-on-crowdpose | Mask R-CNN | #23 | mAP @0.5:0.95: 57.2AP Easy: 69.4AP Medium: 57.9AP Hard: 45.8 |
| panoptic-segmentation-on-cityscapes-val | Mask R-CNN+COCO | #28 | PQth: 54.0 |
| object-detection-on-coco-o | Mask R-CNN
(ResNet-50) | #35 | Average mAP: 17.1 |
| object-detection-on-coco-o | Mask R-CNN (ResNet-50) | #45 | Effective Robustness: -0.11 |
| instance-segmentation-on-coco-test-dev | Mask R-CNN (ResNeXt-101-FPN) | #90 | mask AP: 37.1AP50: 60.0AP75: 39.4APS: 16.9APM: 39.9APL: 53.5 |
| human-pose-keypoint-estimation-on-coco-test-dev | Mask-RCNN | #99 | AP: 63.1AP50: 87.3AP75: 68.7 |
| object-detection-on-coco-test-dev | Mask R-CNN (ResNeXt-101-FPN) | #176 | box mAP: 39.8AP50: 62.3AP75: 43.4APS: 22.1APM: 43.2APL: 51.2 |
| object-detection-on-coco-test-dev | Mask R-CNN (ResNet-101-FPN) | #185 | box mAP: 38.2AP50: 60.3AP75: 41.7APS: 20.1APM: 41.1APL: 50.2 |
| object-detection-on-coco-val2017 | Mask R-CNN (ResNeXt-101-FPN) | #430 | box AP: 36.7AP50: 59.5AP75: 38.9 |