With my interest in GANs soaring as of recently, I have been curious about how certain image processing algorithms work, and whether I could implement them. So, finally, I got around actually trying, practicing, experimenting and in general having fun with the aforementioned algorithms. This repo will contain my experiments with these algorithms, including some relevant tasks I had to do for university.

The file imgplayround.py includes my implementation of the Canny edge detection algorithm (https://en.wikipedia.org/wiki/Canny_edge_detector), with the dependencies being numpy (trigonometry, operations on matrices etc.) PIL (opening files, converting back and forth between images and arrays, showing the images) and scipy (convolutions). The function does a reasonably good job in detecting edges in non incredibly noisy images. However, the implementation will be further improved by experimenting with possible optimization techniques.

Before edge detection:

After edge detection: