Method Details

Details for method 'Panoptic-DeepLab w/ SWideRNet [Mapillary Vistas + Pseudo-labels]'


Method overview

name Panoptic-DeepLab w/ SWideRNet [Mapillary Vistas + Pseudo-labels]
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. Following Naive-Student, this model is additionally trained with pseudo-labels generated from Cityscapes Video and train-extra set (i.e., the coarse annotations are not used, but the images are).
publication Scaling Wide Residual Networks for Panoptic Segmentation
Liang-Chieh Chen, Huiyu Wang, Siyuan Qiao
project page / code
used Cityscapes data fine annotations, video
used external data ImageNet, Mapillary Vistas Research Edition. Cityscapes train-extra set (coarse labels are not used but only images).
runtime n/a
subsampling no
submission date January, 2021
previous submissions


Average results

Metric Value
AP 43.4345
AP50% 68.7408
AP100m 58.8624
AP50m 60.9078


Class results

Class AP AP50% AP100m AP50m
person 39.3022 71.8551 58.628 58.8941
rider 34.9208 68.9528 50.7213 51.1363
car 59.5522 83.0367 79.8055 82.3154
truck 47.8656 59.3632 62.7445 68.4163
bus 57.3514 72.3257 77.6346 85.059
train 45.8594 66.9826 56.0178 55.3642
motorcycle 35.8183 67.241 45.5326 46.2163
bicycle 26.8063 60.1694 39.8146 39.8609



Download results as .csv file

Benchmark page