SiamDW-based trackers for VOT-2019 challenge

We are hiring talented interns: houwen.peng@microsoft.com

SiamDW_D

If you failed to install and run this tracker, please email me (v-hongyy@microsoft.com).

Prerequisites

Install python packages

The python version should be 3.6.x, if not please run,

conda install -y python=3.6

and then run,

bash install.sh

Download pretrained model

Download models here. Unzip networks.zip in SiamDW_D/test/

Set pretrained model path

• row of 14th in SiamDW_D/test/settings/resnext_far/resnext.py, and row of 14th in SiamDW_D/test/settings/senet_far/senet.py please modify main_path to your code SiamDW_D/test directory. eg.

main_path = '/home/hongyuan/VOT2019/SiamDW_D/test/'

Prepare data

You can creat a soft link in test dir. eg.

ln -s $Your-RGBD-data-path data 

or just move RGBD data to SiamDW_D/test/data. Then, modify self.rgbd_path in SiamDW_D/test/settings/envs.py to your RGBD data path.

Prepare to run

Define your experiments in SiamDW_D/test/settings/exp.py. The default is RGBD.

Run tracker

Set gpus id and processes in SiamDW_D/test/parallel_test.py. Then run it.

python parallel_test.py

SiamDW_T

If you failed run this tracker, please be free to email me (v-zhipz@microsoft.com).

Install python packages

sh install.sh

Download model

Download pre-trained model and put it in SiamDW_T/snapshot

Prepare testing data

cd SiamDW_T
mkdir dataset
cd dataset
ln -sfb $you_rgbt_vot_workspace/sequences VOT2019RGBT
cd ..

Run

python rgbt_tracking/test_rgbt.py

SiamDW_LT

If you failed to install and run this tracker, please email me (lzuqer@gmail.com).

Prerequisites

Install python packages

The python version should be 3.6.x, if not please run,

conda install -y python=3.6

and then run,

bash install.sh

Download pretrained model

Download models here. Unzip networks.zip in SiamDW_LT/test/

Set pretrained model path

• row of 14th in SiamDW_LT/test/settings/resnext_far/resnext.py, please modify main_path to your code SiamDW_LT/test directory. eg.

main_path = '/home/v-had/VOT2019/SiamDW_LT/test/'

Prepare data

You can creat a soft link in test dir. eg.

ln -s $Your-LongTerm-data-path data

or just move LongTerm data to SiamDW_LT/test/data. Then, modify self.votlt19_path in SiamDW_LT/test/settings/envs.py to your LongTerm data path.

Prepare to run

Define your experiments in SiamDW_LT/test/settings/exp.py. The default is LongTerm.

Run tracker

Set gpus id and processes in SiamDW_LT/test/parallel_test.py. Then run it.

python parallel_test.py