EmbedMask: Embedding Coupling for One-stage Instance Segmentation

4 Dec 2019Hui YingZhaojin HuangShu LiuTianjia ShaoKun Zhou

Current instance segmentation methods can be categorized into segmentation-based methods that segment first then do clustering, and proposal-based methods that detect first then predict masks for each instance proposal using repooling. In this work, we propose a one-stage method, named EmbedMask, that unifies both methods by taking advantages of them... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Instance Segmentation COCO test-dev EmbedMask(R-101-FPN) mask AP 37,7 # 27
AP50 59.1 # 13
AP75 40.3 # 10
APS 17.9 # 12
APM 40.4 # 11
Instance Segmentation COCO test-dev EmbedMask (ResNet-101-FPN) mask AP 37.7% # 18
AP50 59.1% # 13
AP75 40.3% # 10
APS 17.9% # 12
APM 40.4% # 11
APL 53% # 10

Methods used in the Paper