Compress images with PCA 🎱

MATH 596: Computational Data Science by Erik S. Van Vleck was quite a trip. The philosophy of the course was "sink-or-swim", which means that you either succeed in the class and what’s expected or you just fail it completely. Never have I ever gone through a more brutal multivariate statistics crash course. I must admit that I met some wonderful people there that I never would have had a chance to make an acquaintance with, if it weren’t for this class.

Our first small project was applying Principal Component Analysis (PCA) to some old or new problems. I liked the theory and wanted to see how well I can compress images, therefore extracting the most important (principal, haha) components of a data set and present an approximation of it by only using a fraction of the original set. I did in in python, I recommend running lenna.py if you’re curious to see how it works.

-> Go to the PCA image compression code files