Details for method 'depthAwareSeg_RNN_ff'
Method overview
| name |
depthAwareSeg_RNN_ff |
| challenge |
pixel-level semantic labeling |
| details |
training with fine-annotated training images only (val set is not used); flip-augmentation only in training; single GPU for train&test; softmax loss; resnet101 as front end; multiscale test. |
| publication |
Anonymous
|
| project page / code |
http://www.ics.uci.edu/~skong2/recurrentDepthSeg
|
| used Cityscapes data |
fine annotations |
| used external data |
ImageNet |
| runtime |
n/a |
| subsampling |
no |
| submission date |
March, 2017 |
| previous submissions |
|
Average results
| Metric |
Value |
| IoU Classes | 78.2352 |
| iIoU Classes | 55.9771 |
| IoU Categories | 89.7203 |
| iIoU Categories | 76.9252 |
Class results
| Class |
IoU |
iIoU |
| road | 98.5006 | - |
| sidewalk | 85.4401 | - |
| building | 92.5155 | - |
| wall | 54.4164 | - |
| fence | 60.9183 | - |
| pole | 60.1707 | - |
| traffic light | 72.311 | - |
| traffic sign | 76.8246 | - |
| vegetation | 93.1 | - |
| terrain | 71.5898 | - |
| sky | 94.8327 | - |
| person | 85.2329 | 66.2635 |
| rider | 68.9675 | 46.7332 |
| car | 95.709 | 88.447 |
| truck | 70.115 | 37.3346 |
| bus | 86.5428 | 50.674 |
| train | 75.4961 | 52.0445 |
| motorcycle | 68.3083 | 45.4707 |
| bicycle | 75.4768 | 60.8493 |
Category results
| Category |
IoU |
iIoU |
| flat | 98.5952 | - |
| nature | 92.7961 | - |
| object | 68.29 | - |
| sky | 94.8327 | - |
| construction | 92.9705 | - |
| human | 85.5136 | 67.3694 |
| vehicle | 95.0438 | 86.481 |
Links
Download results as .csv file
Benchmark page
