In this project, we try to detect the motion of an object in an image using Lucas Kanade method. We compute dense flow field vector using Hierarchial Luca & Kanade. The first part of this project deals with Optical flow using Lucas Kanade algorithm. We find that for for large shifts, lucas & Kanade fails to record the movement values accurately. In order to ovecome this limitation, we used Hierarchial Lucas-Kanade method. Later, we used this implementation for Frame interpolations as well.
1) run experiment.py
##Small shifts demo:
Interpolation of intermediate frames from 3 input frames:
