Skip to content

Discriminative and Robust Online Learning for Siamese Visual Tracking (AAAI 2020)

License

Notifications You must be signed in to change notification settings

shallowtoil/DROL

Repository files navigation

DROL

This is the repo for paper "Discriminative and Robust Online Learning for Siamese Visual Tracking" [paper] [results], presented as poster at AAAI 2020.

Introduction

The proposed Discriminative and Robust Online Learning (DROL) module is designed to work with a variety of off-the-shelf siamese trackers. Our method is extensively evaluated over serveral mainstream benchmarks and is believed to induce a consistant performance gain over the given baseline. The model includes but not limited to, as paper evaluated:

Model Zoo

The corresponding offline-trained models are availabe at PySOT Model Zoo.

Get Started

Installation

  • Please find installation instructions for PyTorch and PySOT in INSTALL.md.
  • Add DROL to your PYTHONPATH
export PYTHONPATH=/path/to/drol:$PYTHONPATH

Download models

Download models in PySOT Model Zoo and put the model.pth to the corresponding directory in experiment.

Test tracker

cd experiments/siamrpn_r50_l234_dwxcorr
python -u ../../tools/test.py 	\
	--snapshot model.pth 	\ # model path
	--dataset VOT2018 	\ # dataset name
	--config config.yaml	  # config file

Eval tracker

assume still in experiments/siamrpn_r50_l234_dwxcorr_8gpu

python ../../tools/eval.py 	 \
	--tracker_path ./results \ # result path
	--dataset VOT2018        \ # dataset name
	--num 1 		 \ # number thread to eval
	--tracker_prefix 'model'   # tracker_name

Others

  • For DROL-RPN, we have seperate config file thus each own experiment file folder for vot/votlt/otb/others, where vot is used for VOT-20XX-baseline benchmark, votlt for VOT-20XX-longterm benchmark, otb for OTB2013/15 benchmark, and others is default setting thus for all the other benchmarks, including but not limited to LaSOT/TrackingNet/UAV123.

  • For DROL-FC/DROL-Mask, only experiments on vot/otb are evaluated as described in the paper. Similar to the repo of PySOT, we use config file for vot as default setting.

  • Since this repo is a grown-up modification of PySOT, we recommend to refer to PySOT for more technical issues.

References

Ackowledgement

About

Discriminative and Robust Online Learning for Siamese Visual Tracking (AAAI 2020)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages