Method Details
Details for method 'InstanceCut'
Method overview
| name | InstanceCut |
| challenge | instance-level semantic labeling |
| details | InstanceCut represents the problem by two output modalities: (i) an instance-agnostic semantic segmentation and (ii) all instance-boundaries. The former is computed from a standard CNN for semantic segmentation, and the latter is derived from a new instance-aware edge detection model. To reason globally about the optimal partitioning of an image into instances, we combine these two modalities into a novel MultiCut formulation. |
| publication | InstanceCut: from Edges to Instances with MultiCut A. Kirillov, E. Levinkov, B. Andres, B. Savchynskyy, C. Rother Computer Vision and Pattern Recognition (CVPR) 2017 https://arxiv.org/abs/1611.08272 |
| project page / code | |
| used Cityscapes data | fine annotations, coarse annotations |
| used external data | ImageNet |
| runtime | n/a |
| subsampling | no |
| submission date | November, 2016 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| AP | 13.0479 |
| AP50% | 27.8728 |
| AP100m | 22.0659 |
| AP50m | 26.0512 |
Class results
| Class | AP | AP50% | AP100m | AP50m |
|---|---|---|---|---|
| person | 9.97236 | 27.9935 | 19.669 | 20.1027 |
| rider | 7.99003 | 26.7522 | 13.998 | 14.5724 |
| car | 23.6693 | 44.7985 | 38.8993 | 42.5498 |
| truck | 13.9826 | 22.1945 | 24.8467 | 32.3154 |
| bus | 19.532 | 30.3757 | 34.3696 | 44.7018 |
| train | 15.2266 | 30.1051 | 23.1242 | 31.6875 |
| motorcycle | 9.30228 | 25.0896 | 13.6648 | 14.2706 |
| bicycle | 4.70791 | 15.6732 | 7.95514 | 8.20897 |