| Benchmark | Model | Rank | Results |
|---|---|---|---|
| pedestrian-detection-on-caltech | FRCNN+FPN-Res50+refined feature map+Crowdhuman | #6 | Reasonable Miss Rate: 3.46 |
| pedestrian-detection-on-citypersons | FRCNN+FPN-Res50+refined feature map+Crowdhuman | #8 | Reasonable MR^-2: 10.67 |
| object-detection-on-crowdhuman-full-body | Faster RCNN (ResNet50) | #13 | AP: 84.95mMR: 50.49 |