This project contains a collection of visual tracking algorithms implemented in C++ and/or C. The idea of the project is to help you create, unify, compare and publish your visual tracking algorithms. We also include public tracking datasets for easy integration and analysis of the algorithms.
This project does NOT evaluate and/or rank tracking methods. Many benchmarks with different evaluation methodologies for tracking algorithms are public available. Among the most relevant benchmarks you could find the Visual Tracker Benchmark and the VOT Challenges. We are just interested in collecting and publishing the algorithm implementations and datasets. This way, researches and developers can create their own trackers, compare and execute different algorithms using their own sequences or existing annotated datasets.
We provide a cross-platform framework for easy integration and execution of tracking algorithms on annotated datasets. The project uses cmake system generator to facilitate the development and testing of the algorithms in your own platform/IDE.
The vivaTracker framework is a C++11 cross-platform project to create, compare and test object tracking algorithms using OpenCV 3.x.x.
For more detailed information about the implementations and datasets provided in this project reefer to the project's list.
- CMake > 3.0
- C++11 compiler
- OpenCV > 3.0
- Eigen Library > 3.0
Mac, Linux, Windows
For more details and documentations refer to the project's wiki.
This project was supported by the VIVALab, University of Ottawa.
Licensing information can be found inside the license folder for each of the project component and trackers.