Apologies if I missed this – I read the FAQ and sidebar, but it wasn’t exactly what I was looking for.

Some context:

I’m a self-taught engineer who learned CS in my 30s through brute forcing Berkeley and MIT’s online curriculum. I have a regular SWE job and can do regular SWE things.

I’m fascinated by Computer Vision and would love to dabble in it (even if I don’t move my career in that direction). In high school, I was always fairly decent at math, but I haven’t touched math since I was 16 (I took Calculus 1 in high school and passed the AP exam so I didn’t have to do it in college since I got a useless liberal arts degree). That being said, it’s been over 20 years since I’ve looked at math seriously (outside of the basic math stuff I do at work).

Question:

What is your recommendation on how to approach learning (or re-learning) math so I can better understand Computer Vision? I’m not in a rush and genuinely enjoy learning hard stuff the right way, but I’d like to minimize the amount of time before I can become useful in CV.

I know what I need to know (multivar calc, linear algebra at the minimum), but my challenge is that I’m not sure how far back to start. I’ve definitely forgotten what derivatives are, so I at least need to start with calc, but then I probably have forgotten a ton of geometry, precalc, and even reg algebra concepts.

So I think my approach needs to be “refill the gaps in my memory from what I learned in school” and then “learn the new stuff”.

Would love any recommendations on:

How can I assess where my gaps in foundational math are?

Order of operations when it comes to learning this stuff

I’m leaning towards not taking classes, but getting good textbooks and working through them. Just need to narrow down the scope of books and the order of the books. Thanks for your help! 🙂