Module1 - final - A4 - KLT Tracker

You will use the widely used Kanade-Lucas-Tomasi tracker to track a region-of-interest (ROI) in a streaming video. You start by implementing the tracker from scratch and then compare with OpenCV’s implementation. To start with, the ROI can be selected manually. You can then investigate how you can use various ways to automate the initialisation of the tracker, e.g. by making use of some of the material from the deep learning page or to investigate more advanced and state of the art tracking method, such as those presented by Michael Felsberg. 

Tools you will use: OpenCV, optionally TensorFlow, ROS, etc
Background: All are good to go!

The detailed instructions can be found in this file wasp_as1_2018_tracking_assignment.pdf Download wasp_as1_2018_tracking_assignment.pdf

Please see these instructions for information about what to submit and about the peer-review.

The test images (view0.png Download view0.png and view1.png Download view1.png) mentioned in the assignment can be seen below