================ TLDC ================ OpenTLDC is an algorithm for tracking of unknown objects in unconstrained video streams. It is based on TLD, published by Zdenek Kalal. It does not depend on any training data, it learns and detects objects in real time. This package includes a shared library and a stand-alone application. ================ License ================ TLDC is released under the GPL license version 3.0. For more info see http://www.gnu.org/licenses/gpl-3.0.txt ================ Installing ================ 1.) TLDC requires the following libraries: - OpenCV 2.1 (http://opencv.willowgarage.com/wiki/) - Eigen 3.0.2 (http://eigen.tuxfamily.org/index.php?title=Main_Page) 2.) CMake is used for compilation. Building can take the form: cd path/to/tldc mkdir build cd build cmake .. make 3.) To install: [sudo] make install To set a different install prefix, change the 'cmake ..' step to: cmake -DCMAKE_INSTALL_PREFIX=/your/prefix/ .. and then make install. ================ Usage (stand-alone application) ================ tldc [args] [-x] : Initial boundingbox x-positions, expects two values (xmin xmax) [-y] : Initial boundingbox y-positions, expects two values (ymin ymax) [-vid] : Uses videosource, expects path to source ("/home/.....") [-cam] : Uses camera as source, expects index of camera [-nodisplay] : Turns output window off [-fr] : First frame index of video stream You have to set [-x] and [-y] coordinates to apply the initial bounding box that should be learned and tracked. Example: ./tldc -vid "/home/tldc/test.avi" -fr 1061 -x 310 380 -y 110 200 ================ Related Work ================ TLDC is a rewrite of TLD in C++, originally written in MATLAB by Zdenek Kalal. For more info, see: @article{Kalal2010, author = {Kalal, Z and Matas, J and Mikolajczyk, K}, journal = {Conference on Computer Vision and Pattern Recognition}, title = {{P-N Learning: Bootstrapping Binary Classifiers by Structural Constraints}}, year = {2010} } ================ Acknowledgements ================ TLDC was started as a student project under the supersivion of Arne Kreutzmann , Jasper van de Ven, Dr.-Ing. Falko Schmid and Prof. Ph. D. Christian Freksa, of the workgroup Cognitive Systems at the University of Bremen. Contact: Cognitive Systems (CoSy) FB3 - Informatics Universität Bremen P.O. Box 330 440 28334 Bremen cosy@informatik.uni-bremen.de ================ (c) 2011 Sascha Schrader, Stefan Brending, Adrian Block