Method Details
Details for method 'SQ'
Method overview
| name | SQ |
| challenge | pixel-level semantic labeling |
| details | |
| publication | Speeding up Semantic Segmentation for Autonomous Driving Michael Treml, José Arjona-Medina, Thomas Unterthiner, Rupesh Durgesh, Felix Friedmann, Peter Schuberth, Andreas Mayr, Martin Heusel, Markus Hofmarcher, Michael Widrich, Bernhard Nessler, Sepp Hochreiter NIPS 2016 Workshop - MLITS Machine Learning for Intelligent Transportation Systems Neural Information Processing Systems 2016, Barcelona, Spain https://openreview.net/pdf?id=S1uHiFyyg |
| project page / code | |
| used Cityscapes data | fine annotations |
| used external data | ImageNet |
| runtime | 0.06 s Nvidia TX1 for 480 x 320 |
| subsampling | no |
| submission date | October, 2016 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 59.8406 |
| iIoU Classes | 32.2562 |
| IoU Categories | 84.3166 |
| iIoU Categories | 65.9716 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 96.9235 | - |
| sidewalk | 75.3709 | - |
| building | 87.8555 | - |
| wall | 31.5868 | - |
| fence | 35.6884 | - |
| pole | 50.9243 | - |
| traffic light | 52.0044 | - |
| traffic sign | 61.659 | - |
| vegetation | 90.8987 | - |
| terrain | 65.7836 | - |
| sky | 93.0496 | - |
| person | 73.8124 | 47.834 |
| rider | 42.5881 | 21.7396 |
| car | 91.4894 | 84.6342 |
| truck | 18.8325 | 7.76417 |
| bus | 41.247 | 20.991 |
| train | 33.3277 | 16.6256 |
| motorcycle | 34.0351 | 14.8569 |
| bicycle | 59.8937 | 43.6044 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 96.6666 | - |
| nature | 90.431 | - |
| object | 57.0409 | - |
| sky | 93.0496 | - |
| construction | 87.4571 | - |
| human | 75.6384 | 49.9549 |
| vehicle | 89.9324 | 81.9882 |