Computer Vision (31/07/2021)
Computer vision aims to extract information from the surroundings using cameras. This is much harder than you might think - think about how you might tell a computer how to recognise an elephant for example:
![]() |
| Happy the elephant at Bronx Zoo |
You couldn't just say that it is grey, has four legs and a trunk. The computer probably doesn't know what an animal even is or what any of those words mean. Sure it might be easy to program it to look for grey pixels, but then it might mistake a rock for an elephant. And what if the camera is positioned in a way that makes it difficult to distinguish the different legs, or if the trunk wasn't visible?
I have been interested in computer vision for quite a while now but had no idea where to start. Recently, I got this book Concise Computer Vision: An Introduction Into Theory and Algorithms. It's got good reviews, but I had trouble implementing the stuff from the first few chapters.
A better resource for beginners might be MIT's Introduction to Computational Thinking (2021). It uses a new programming language called Julia but the ideas should transfer to any language.
The previous year's version of the course, featuring Grant Sanderson from 3Blue1Brown, is available as a youtube playlist (here).

Comments
Post a Comment