Method Details
Details for method 'StixelPointNet'
Method overview
| name | StixelPointNet |
| challenge | instance-level semantic labeling |
| details | An adapted version of the PointNet is trained on Stixels as input for instance segmentation. |
| publication | Learning Stixel-Based Instance Segmentation Monty Santarossa, Lukas Schneider, Claudius Zelenka, Lars Schmarje, Reinhard Koch, Uwe Franke IV 2021 |
| project page / code | |
| used Cityscapes data | fine annotations |
| used external data | |
| runtime | 0.035 s Intel Core i7-4790K CPU, NVIDIA GeForce GTX 1070 GPU |
| subsampling | no |
| submission date | February, 2020 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| AP | 8.5366 |
| AP50% | 19.2537 |
| AP100m | 15.1415 |
| AP50m | 18.2868 |
Class results
| Class | AP | AP50% | AP100m | AP50m |
|---|---|---|---|---|
| person | 9.04899 | 24.6467 | 18.6144 | 19.2545 |
| rider | 7.30913 | 24.8968 | 13.1433 | 13.7667 |
| car | 15.8422 | 31.8306 | 26.5018 | 29.5472 |
| truck | 12.7755 | 22.304 | 21.41 | 27.6841 |
| bus | 16.3371 | 27.3392 | 29.3031 | 42.8191 |
| train | 0 | 0 | 0 | 0 |
| motorcycle | 3.52509 | 11.5937 | 5.63018 | 6.34613 |
| bicycle | 3.45483 | 11.4185 | 6.52891 | 6.87639 |