Skip to content

This is the add-on package for the pysot project. SiamMask_E is an updated version of SiamMask with ellipse fitting and bounding box scale refinement.

License

Notifications You must be signed in to change notification settings

baoxinchen/siammask_e

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

29 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PWC PWC PWC

SiamMask_E

In this project, we demonstrate a novel algorithm that uses ellipse fitting to estimate the bounding box rotation angle and size with the segmentation(mask) on the target for online and real-time visual object tracking. Our method, SiamMask E, improves the bounding box fitting procedure of the state-of-the-art object tracking algorithm SiamMask and still retains a fast-tracking frame rate (80 fps) on a system equipped with GPU (GeForce GTX 1080 Ti or higher). We tested our approach on the visual object tracking datasets (VOT2016, VOT2018, and VOT2019) that were labeled with rotated bounding boxes. By comparing with the original SiamMask, we achieved an improved Accuracy of 65.2% and 30.9% EAO on VOT2019, which is 5.6% and 2.6% higher than the original SiamMask.

This repository is the add-on package for PySOT project.

Paper

Fast Visual Object Tracking with Rotated Bounding Boxes

News

Aug. 14, 2019: posted results on datasets: VOT2016, VOT2018, VOT2019.

Aug. 12, 2019: accepted as a workshop paper in ICCV2019 The Visual Object Tracking Challenge Workshop VOT2019.

Merge file to pysot

cd siammask_e
bash install.sh [/path/to/pysot]

Webcam demo

python tools/demo.py \
    --config experiments/siammaske_r50_l3/config.yaml \
    --snapshot experiments/siammaske_r50_l3/model.pth \
    # --video demo/bag.avi # (in case you don't have webcam)

Model

Please use the same model as SiamMask, refer to PySOT Model Zoo

Short-term Tracking on VOT2016, 2018, 2019

Sample outputs

Reference

@article{chen2019fast,
title={Fast Visual Object Tracking with Rotated Bounding Boxes},
author={Chen, Bao Xin and Tsotsos, John K},
journal={arXiv preprint arXiv:1907.03892},
year={2019}
}

@article{wang2018fast,
title={Fast Online Object Tracking and Segmentation: A Unifying Approach},
author={Wang, Qiang and Zhang, Li and Bertinetto, Luca and Hu, Weiming and Torr, Philip HS},
journal={arXiv preprint arXiv:1812.05050},
year={2018}
}

@article{li2018siamrpn++,
title={SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks},
author={Li, Bo and Wu, Wei and Wang, Qiang and Zhang, Fangyi and Xing, Junliang and Yan, Junjie},
journal={arXiv preprint arXiv:1812.11703},
year={2018}
}

About

This is the add-on package for the pysot project. SiamMask_E is an updated version of SiamMask with ellipse fitting and bounding box scale refinement.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published