Method Details

Details for method 'Panoptic-DeepLab w/ SWideRNet [Cityscapes-fine]'


Method overview

name Panoptic-DeepLab w/ SWideRNet [Cityscapes-fine]
challenge instance-level semantic labeling
details We revisit the architecture design of Wide Residual Networks. We design a baseline model by incorporating the simple and effective Squeeze-and-Excitation and Switchable Atrous Convolution to the Wide-ResNets. Its network capacity is further scaled up or down by adjusting the width (i.e., channel size) and depth (i.e., number of layers), resulting in a family of SWideRNets (short for Scaling Wide Residual Networks). We demonstrate that such a simple scaling scheme, coupled with grid search, identifies several SWideRNets that significantly advance state-of-the-art performance on panoptic segmentation datasets in both the fast model regime and strong model regime.
publication Scaling Wide Residual Networks for Panoptic Segmentation
Liang-Chieh Chen, Huiyu Wang, Siyuan Qiao
project page / code
used Cityscapes data fine annotations
used external data ImageNet
runtime n/a
subsampling no
submission date October, 2020
previous submissions


Average results

Metric Value
AP 37.9881
AP50% 61.0086
AP100m 53.6669
AP50m 55.3675


Class results

Class AP AP50% AP100m AP50m
person 36.7512 67.5584 56.1119 56.1588
rider 33.1914 66.0443 48.3469 48.7642
car 57.2325 80.6131 77.6682 80.085
truck 38.8198 49.4483 51.8594 57.2318
bus 45.0164 56.7782 67.0208 73.1623
train 38.8936 54.7664 53.0612 51.2623
motorcycle 30.1841 58.538 39.747 40.6594
bicycle 23.8155 54.3223 35.5198 35.6158



Download results as .csv file

Benchmark page