This homework consists of 3 major sections.
In the first section, you will implement a simple Lucas-Kanade (LK) tracker with one single template.
The second section requires you to implement a motion subtraction method for tracking moving pixels in a scene.
In the final section, you shall study efficient tracking such as inverse composition.
Using git pull
from the terminal while inside the git folder should help download all of the files.
Please download the data
folder from this Drive link. Please let us know early if you are facing issues in downloading. The internal folder structure should look something like this
hw2
>code
>>...
>data
>>...
>16_820_F23_HW2_Lucas_Kanade_Tracking.pdf
>README.md
Visualization: You are free to visualize the bounding boxes and frames any way you see fit. To help with questions 1.3 and 1.4
we have provided a helper script plotRects.py
which you can use.
Example usage: python plotRects.py q1.3 car