Computer Vision Algorithms

February, 2016

Overview

As part of a computer vision course that I took during my masters the weekly homeworks were about understanding image processing algorithms and writing the code to implement them without using any preexisting computer vision libraries.

Software

I chose to write my software in C++ which can all be found here.

The functions implemented were:

Identifying Connected Regions
Morphological Operations
Histogram Equalization and Lighting Correction
Colour Segmentation
Canny Edge Detector
Hough Transform

Image Results of Tested Algorithms

Identifying connected regions and colouring each segment differently.

Using morphological operations to remove noise and fill in broken regions.

Histogram equalization and lighting correction.

Colour segmentation by training with an image of only skin.

Canny Edge Detection

Using the hough Transform to find lines that fit the edges.