Method Details
Details for method 'Sem2Ins'
Method overview
name | Sem2Ins |
challenge | panoptic semantic labeling |
details | Anonymous NeurIPS19 submission #4671 |
publication | Anonymous |
project page / code | |
used Cityscapes data | fine annotations |
used external data | |
runtime | n/a |
subsampling | no |
submission date | July, 2019 |
previous submissions |
Average results
Metric | All | Things | Stuff |
---|---|---|---|
PQ | 52.2828 | 37.7785 | 62.8313 |
SQ | 78.9256 | 77.2318 | 80.1574 |
RQ | 65.1681 | 48.93 | 76.9777 |
Class results
Class | PQ | SQ | RQ |
---|---|---|---|
road | 88.9841 | 89.1016 | 99.8682 |
sidewalk | 74.2366 | 82.7333 | 89.7301 |
building | 86.0132 | 88.9198 | 96.7312 |
wall | 30.469 | 75.0172 | 40.616 |
fence | 35.5391 | 75.3354 | 47.1744 |
pole | 51.9367 | 67.2766 | 77.1988 |
traffic light | 47.7053 | 71.7938 | 66.4477 |
traffic sign | 66.8669 | 76.3548 | 87.574 |
vegetation | 88.2794 | 89.8962 | 98.2016 |
terrain | 37.2929 | 76.6719 | 48.6396 |
sky | 83.8211 | 88.6311 | 94.5731 |
person | 39.7536 | 73.6291 | 53.9916 |
rider | 39.6643 | 71.7439 | 55.2859 |
car | 47.7202 | 80.0152 | 59.6389 |
truck | 36.8365 | 85.3378 | 43.1655 |
bus | 43.2005 | 85.4289 | 50.5689 |
train | 31.6064 | 78.6798 | 40.1709 |
motorcycle | 33.9483 | 73.6321 | 46.1053 |
bicycle | 29.4986 | 69.3874 | 42.5128 |