Fast Online Object Tracking and Segmentation: A Unifying Approach

CVPR 2019 Qiang WangLi ZhangLuca BertinettoWeiming HuPhilip H. S. Torr

In this paper we illustrate how to perform both visual object tracking and semi-supervised video object segmentation, in real-time, with a single simple approach. Our method, dubbed SiamMask, improves the offline training procedure of popular fully-convolutional Siamese approaches for object tracking by augmenting their loss with a binary segmentation task... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Visual Object Tracking DAVIS 2016 SiamMask Jaccard (Mean) 71.7 # 9
Visual Object Tracking DAVIS 2016 SiamMask Jaccard (Recall) 86.8 # 9
Visual Object Tracking DAVIS 2016 SiamMask Jaccard (Decay) 3.0 # 2
Visual Object Tracking DAVIS 2016 SiamMask F-measure (Mean) 67.8 # 9
Visual Object Tracking DAVIS 2016 SiamMask F-measure (Recall) 79.8 # 9
Visual Object Tracking DAVIS 2016 SiamMask F-measure (Decay) 2.1 # 1
Visual Object Tracking DAVIS-2017 SiamMask Jaccard (Mean) 54.3 # 4
Visual Object Tracking DAVIS-2017 SiamMask Jaccard (Recall) 62.8 # 3
Visual Object Tracking DAVIS-2017 SiamMask Jaccard (Decay) 19.3 # 4
Visual Object Tracking DAVIS-2017 SiamMask F-measure (Mean) 58.5 # 4
Visual Object Tracking DAVIS-2017 SiamMask F-measure (Recall) 67.5 # 4
Visual Object Tracking DAVIS-2017 SiamMask F-measure (Decay) 20.9 # 2
Visual Object Tracking VOT2017/18 SiamMask Expected Average Overlap (EAO) 0.380 # 6
Visual Object Tracking YouTube-VOS SiamMask Jaccard (Seen) 54.3 # 4
Visual Object Tracking YouTube-VOS SiamMask O (Average of Measures) 52.8 # 3
Visual Object Tracking YouTube-VOS SiamMask Jaccard (Unseen) 45.1 # 3
Visual Object Tracking YouTube-VOS SiamMask F-Measure (Seen) 58.2 # 4
Visual Object Tracking YouTube-VOS SiamMask F-Measure (Unseen) 47.7 # 3