COL780-Computer Vision

Programming Assignments of Computer Vision Course at IIT Delhi.

Implemented various concepts of Computer Vision:

Assiginment1:

Background Subtraction in these scenario:

1. Baseline,
2. Illumination changes,
3. Camera Shake(jitter),
4. Dynamic Scene(changing background) and
5. Pan-tilt-zoom videos.

Assiginment2:

Template Tracking:

1. Part 1 Appearance-based Techniques: sum of squared distance (SSD) and normalised cross correlation (NCC).
2. Part 2 Lucas Kanade Algorithm
3. Part 3 Pyramid based Lucas Kanade Algorithm
4. Part 4 Live Demo

Assiginment3:

Pedestrian Detection using:

1. HoG(Histogram of Oriented Gradients):
   1. pretrained HoG detector,
   2. train your own classifier by extracting HoG features.
2. Faster-RCNN.