C++ code for template-based tracking method, shortened as ESM, proposed in Homography-based 2D Visual Tracking and Servoing. ESM method elimites the need to compute the Hessian matrix of the input images. Thus it significantly improves the speed to 75 frames per second (8G RAM, Intel I7).

1. Classes and methods from OpenCV 2.4.12 are used in the code. OpenCV 2.4.12 can be downloaded from https://opencv.org/. The .props file included in the folder must be imported before you run the code.
2. Debug and Win64 should be choosed. If you want to run in Release mode, you need to create and include a new .props file.

1. Gregory Hager and Peter Belhumeur. Efficient Region Tracking With Parametric Models of Geometry and Illumination. IEEE Trans. on Pattern Analysis and Machine Intelligence, 1998.
2. Selim Benhimane and Ezio Malis. Real-time image-based tracking of planesusing Efficient Second-order Minimization. In Proc IEEE Conference on Intelligent Robotics and Systems, 2004.
3. Selim Benhimane and Ezio Malis. Homography-based 2D Visual Tracking and Servoing. International Journal of Robotics Research, 2007.
4. Rogerio Richa, Raphael Sznitman and Gregory Hager. Robust Similarity Measures for Gradient-based Direct Visual Tracking. CIRL Technical Report, 2012

The code was written a while ago and I was not good at programming at that time. You might find the code confusing due to lack of comments.