Method Details
Details for method 'ENet'
Method overview
| name | ENet |
| challenge | pixel-level semantic labeling |
| details | |
| publication | ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation Adam Paszke, Abhishek Chaurasia, Sangpil Kim, Eugenio Culurciello https://arxiv.org/abs/1606.02147 |
| project page / code | https://github.com/e-lab/ENet-training |
| used Cityscapes data | fine annotations |
| used external data | |
| runtime | 0.013 s NVIDIA Titan X |
| subsampling | 2 |
| submission date | May, 2016 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| IoU Classes | 58.2878 |
| iIoU Classes | 34.363 |
| IoU Categories | 80.3973 |
| iIoU Categories | 63.9772 |
Class results
| Class | IoU | iIoU |
|---|---|---|
| road | 96.3273 | - |
| sidewalk | 74.2395 | - |
| building | 85.0487 | - |
| wall | 32.1642 | - |
| fence | 33.2327 | - |
| pole | 43.4502 | - |
| traffic light | 34.1022 | - |
| traffic sign | 44.0244 | - |
| vegetation | 88.6077 | - |
| terrain | 61.3903 | - |
| sky | 90.6385 | - |
| person | 65.5102 | 47.6293 |
| rider | 38.4262 | 20.7912 |
| car | 90.5971 | 80.0338 |
| truck | 36.9046 | 17.5274 |
| bus | 50.5119 | 26.8045 |
| train | 48.0834 | 21.8271 |
| motorcycle | 38.8017 | 20.8791 |
| bicycle | 55.4076 | 39.4118 |
Category results
| Category | IoU | iIoU |
|---|---|---|
| flat | 97.3417 | - |
| nature | 88.2815 | - |
| object | 46.7501 | - |
| sky | 90.6385 | - |
| construction | 85.4022 | - |
| human | 65.4968 | 49.2703 |
| vehicle | 88.87 | 78.684 |