README

This will be a C++ implementations of the OpenTLD (aka Predator)
----------------------------------------------------------------------------
This is a work in progress, as of right now the code is functional but pretty slow.
----------------------------------------------------------------------------

Quick Test on Mouse dataset
=============================
bin/run_tld -p parameters.yml -s <path_to_video.avi>

- "r" key to pause the video and add another model
- "left click" on "Examples" GUI to remove existing model

TODO:
- sometimes, there are segfault..
- the model list will not automatically refresh. (may want to add a new model to refresh the "Example" GUI)



Installation Notes
=====================================
git clone git@github.com:alantrrs/OpenTLD.git
cd OpenTLD
mkdir build
cd build
cmake ../src/
make
cd ../bin/
%To run from camera
./run_tld -p ../parameters.yml
%To run from file
./run_tld -p ../parameters.yml -s ../datasets/06_car/car.mpg
%To init bounding box from file
./run_tld -p ../parameters.yml -s ../datasets/06_car/car.mpg -b ../datasets/06_car/init.txt
%To train only in the firs frame (no tracking, no learning)
./run_tld -p ../parameters.yml -s ../datasets/06_car/car.mpg -b ../datasets/06_car/init.txt -no_tl 
%To test the final detector (Repeat the video, first time learns, second time detects)
./run_tld -p ../parameters.yml -s ../datasets/06_car/car.mpg -b ../datasets/06_car/init.txt -r

=====================================
Evaluation
=====================================
The output of the program is a file called bounding_boxes.txt which contains all the detections made through the video. This file should be compared with the ground truth file to evaluate the performance of the algorithm. This is done using a python script:
python ../datasets/evaluate_vis.py ../datasets/06_car/car.mpg bounding_boxes.txt ../datasets/06_car/gt.txt

====================================
Thanks
====================================
To Zdenek Kalal for realeasing his awesome algorithm