Tracker Benchmark

This repository is modified from https://github.com/jwlim/tracker_benchmark.

Goals:

• Light Weight. Extra trackers and results are removed in this repository
• Dynamic Import. Add only one file for a new tracker. Do not need to import in __init__.py
• Show More. Show more results. (e.g.: AUC, Precision)
• No Bug. Bugs are fixed.
• Compatible. Easy to use in python2 and python3

Run new tracker

• Add script to running tracker in scripts/bscripts (see ExampleTracker.py)

Usage

Default (for all trackers, all sequences, all evaltypes(OPE, SRE, TRE))

• command : python run_trackers.py

For specific trackers, sequences, evaltypes

• command : python run_trackers.py -t "tracker" -s "sequence" -e "evaltype"
• sequence can be name of Sequence, 'tb50', 'tb100' and 'cvpr13' (using data/tb_50.txt, tb_100.txt, cvpr13.txt)
• e.g.
  • python run_trackers.py -t IVT,TLD -s Couple,Crossing -e OPE,SRE)
  • python run_trackers.py -s tb50

Libraries

• matplotlib
• numpy
• Python Imaging Library (PIL)

Troubleshooting

Segmentaion Fault when running 'python run_tracker.py ...' on MacOSX

• Set DYLD_LIBRARY_PATH environment variable.
• e.g.: export DYLD_LIBRARY_PATH=/usr/local/Cellar/python/2.7.11/Frameworks/Python.framework/Versions/Current/lib/:$DYLD_LIBRARY_PATH