Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation

22 Nov 2019Bowen ChengMaxwell D. CollinsYukun ZhuTing LiuThomas S. HuangHartwig AdamLiang-Chieh Chen

In this work, we introduce Panoptic-DeepLab, a simple, strong, and fast system for panoptic segmentation, aiming to establish a solid baseline for bottom-up methods that can achieve comparable performance of two-stage methods while yielding fast inference speed. In particular, Panoptic-DeepLab adopts the dual-ASPP and dual-decoder structures specific to semantic, and instance segmentation, respectively... (read more)

PDF Abstract

Results from the Paper


#2 best model for Panoptic Segmentation on Cityscapes test (using extra training data)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
RESULT LEADERBOARD
Panoptic Segmentation Cityscapes test Panoptic-Deeplab PQ 65.5 # 2
Semantic Segmentation Cityscapes test Panoptic-DeepLab Mean IoU (class) 84.2% # 3
Instance Segmentation Cityscapes test Panoptic-DeepLab [Mapillary Vistas] Average Precision 39.0 # 2
Instance Segmentation Cityscapes test Panoptic-DeepLab [Cityscapes-fine] Average Precision 34.6 # 3
Panoptic Segmentation Cityscapes val Panoptic-DeepLab (X71) PQ 64.1 # 3
mIoU 81.5 # 3
AP 38.5 # 4
Semantic Segmentation Cityscapes val Panoptic-DeepLab mIoU 81.5% # 2
Panoptic Segmentation COCO test-dev Panoptic-DeepLab (Xception-71) PQ 41.4 # 7
PQst 35.9 # 3
PQth 45.1 # 7
Panoptic Segmentation Mapillary val Panoptic-DeepLab (X71) PQ 40.3 # 2