Method Details
Details for method 'GAIS-Net'
Method overview
| name | GAIS-Net |
| challenge | instance-level semantic labeling |
| details | Geometry-Aware Instance Segmentation with Disparity Maps |
| publication | Geometry-Aware Instance Segmentation with Disparity Maps Cho-Ying Wu, Xiaoyan Hu, Michael Happold, Qiangeng Xu, Ulrich Neumann Scalability in Autonomous Driving, workshop at CVPR 2020 https://arxiv.org/abs/2006.07802 |
| project page / code | https://github.com/choyingw/GAIS-Net |
| used Cityscapes data | fine annotations, stereo |
| used external data | ImageNet, COCO |
| runtime | n/a |
| subsampling | no |
| submission date | March, 2020 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| AP | 32.262 |
| AP50% | 59.4847 |
| AP100m | 44.5528 |
| AP50m | 46.6115 |
Class results
| Class | AP | AP50% | AP100m | AP50m |
|---|---|---|---|---|
| person | 36.0351 | 68.7388 | 51.8498 | 51.9337 |
| rider | 29.0225 | 66.8026 | 41.3054 | 41.9439 |
| car | 52.8019 | 77.9875 | 71.2109 | 73.4234 |
| truck | 29.6941 | 42.6566 | 39.9653 | 45.3519 |
| bus | 39.7577 | 59.0655 | 54.9122 | 58.3544 |
| train | 28.9372 | 56.0547 | 41.4944 | 45.7776 |
| motorcycle | 23.2985 | 54.9956 | 29.4237 | 29.8348 |
| bicycle | 18.5487 | 49.5764 | 26.2604 | 26.2723 |