Method Details

Details for method 'Panoptic-DeepLab [Cityscapes-fine]'


Method overview

name Panoptic-DeepLab [Cityscapes-fine]
challenge instance-level semantic labeling
details Our proposed bottom-up Panoptic-DeepLab is conceptually simple yet delivers state-of-the-art results. The Panoptic-DeepLab adopts dual-ASPP and dual-decoder modules, specific to semantic segmentation and instance segmentation respectively. The semantic segmentation prediction follows the typical design of any semantic segmentation model (e.g., DeepLab), while the instance segmentation prediction involves a simple instance center regression, where the model learns to predict instance centers as well as the offset from each pixel to its corresponding center. This submission exploits only Cityscapes fine annotations.
publication Panoptic-DeepLab
Bowen Cheng, Maxwell D. Collins, Yukun Zhu, Ting Liu, Thomas S. Huang, Hartwig Adam, Liang-Chieh Chen
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date September, 2019
previous submissions


Average results

Metric Value
AP 33.7867
AP50% 56.1302
AP100m 49.7165
AP50m 52.4989


Class results

Class AP AP50% AP100m AP50m
person 32.7353 61.8275 52.8153 53.0831
rider 28.0997 60.0676 42.964 43.8613
car 52.2102 75.1291 74.3685 77.1542
truck 32.5228 41.6384 47.4379 52.4783
bus 41.1263 53.087 62.8981 69.8654
train 36.6678 55.5963 49.5216 54.8966
motorcycle 25.7379 51.7468 34.5523 35.4885
bicycle 21.1935 49.9488 33.1746 33.1634



Download results as .csv file

Benchmark page