Skip to content

Latest commit

 

History

History
93 lines (71 loc) · 3.51 KB

test.md

File metadata and controls

93 lines (71 loc) · 3.51 KB
Raw

Test

Test on a specific video

eg,

python siamese_tracking/run_video.py --arch SiamRPNRes22 --resume snapshot/CIResNet22_RPN.pth --video videos/bag.mp4
  • The opencv version here is 4.1.0.25, and older versions may be not friendly to some functions.
  • If you try to conduct this project on a specific tracking task, eg. pedestrian tracking, it's suggested that you can tuning hyper-parameters on your collected data with our tuning toolkit detailed below.

Test through webcam

eg,

python siamese_tracking/run_webcam.py --arch SiamRPNRes22 --resume snapshot/CIResNet22_RPN.pth
  • The opencv version here is 4.1.0.25, and older versions may be not friendly to some functions.
  • You can embed any tracker for fun. This is also a good way to design experiments to determine how environmental factors affect your tracker.

Test on benchmarks

Data preparation

The test dataset (OTB or VOT) should be arranged in dataset directory. Your directory tree should look like this:

${Tracking_ROOT}
|—— experimnets
|—— lib
|—— snapshot
|—— dataset
  |—— OTB2013.json
  |—— OTB2015.json
  |—— VOT2015
     | —— videos...
  |—— OTB2013 (or VOT2015...)
     |—— videos...
|—— run_tracker.py
|—— ...

OTB2013.json and OTB2015.json can be download here.

Conda preparation

sh install_fc.sh or sh install_rpn.sh

Toolkit preparation

  • Set up vot-toolkit according to official tutorial
  • Modify path_to/toolkit in lib/core/get_eao.m to your vot-toolkit path
  • In your matlab install path (MATLAB2017b or higher),
cd $matlab_path/R2018b/extern/engines/python
python setup.py install

Run tracker

CUDA_VISIBLE_DEVICES=0 python ./siamese_tracking/test_siamfc.py --arch SiamFCRes22 --resume ./snapshot/CIResNet22.pth --dataset OTB2013
or 
CUDA_VISIBLE_DEVICES=0 python ./siamese_tracking/test_siamrpn.py --arch SiamRPNRes22 --resume ./snapshot/CIResNet22_RPN.pth --dataset VOT2017 --cls_type thinner

Analysz testing results

  • OTB
python ./lib/core/eval_otb.py OTB2013 ./result SiamFC* 0 1
  • VOT
python ./lib/core/eval_vot.py VOT2017 ./result 2>&1 | tee logs/siamfc_eval.log

Attention !!

  • Recently we found that the image is slightly inconsistent while using different OpenCV version. And the speed of some opencv versions are relatively slow for some reason. It is recommended that you install packages above.
  • The SiamRPN based model is trained on pytorch0.4.1, since we found that memory leak happens while testing SiamRPN on pytorch0.3.1 with multithread tools.