Kalman Filter Enhanced Template Tracking Using Normalized Cross Correlation (NCC)

https://is.gd/VEHD9Y TemplateTuning secret sale page.
Template tracking (NCC) results using a Kalman filter to reduce the search window size to increase speed/performance. I worked on this problem as part of my computer vision class at Utah State University (CVPRIP 1). Box meanings are as follows: Purple: the possible top-left locations of the target according to the Kalman filter Yellow: the Kalman filter determined search area Blue: the statistical window (NCC was run on this whole window to get statistical information) Pale green: the possible top-left locations results for the statistical window Red: where the Kalman filter thought the target should be, but it wasn't there according to the threshold Lime Green: the actual matched location of the target.