Where to start learning computer vision?

Question

IT education

Good day )

Having tried myself in popular areas of programming and development, I realized that I don’t like it all. In general, there is no desire to even get involved with all this. I tried at a very primitive level, since I just finished 11th grade and my knowledge is poor. But even a little contact was enough for me to understand: this is not mine.

What 's interesting to me is this:

0. Data mining
1. Computer vision
2. Neural networks and decision-making
systems 4. Probability analysis systems and process prediction.

I shamelessly borrowed the list from our favorite toaster, having previously googled what each direction represents and “crossed out” those of no interest)

By the way, I entered the mathematical direction and there will be very, very little programming.

How to choose one thing when everything is interesting? But this is a rather rhetorical question, depending on personal preferences, and I understand this. So I would like to start with computer vision. Among the skills in vacancies, the following list is constantly found:

Mathematical training:
basics of linear algebra
analytical geometry
probability theory
Algorithmic training:
knowledge of basic algorithms (sorting, search...) and 3D graphics and computer vision
algorithms analysis of the effectiveness of algorithms
knowledge of Python and C++
technical English

Everything on this list is clear to me. Besides

• knowledge of basic algorithms (sorting, search...) and algorithms of 3D graphics and computer
  
• algorithm efficiency analysis
  

Mathematics - I will work at the university, English is by itself, I will teach python and C ++ myself (I expect to learn at least a little in these PLs in 3 years).

And finally, the question itself.

Advise what to read on this topic. There are books, but I'm afraid to make a mistake and choose some old and outdated dry reference book. Unfortunately, I have experience in choosing the most terrible material. Ready to read any book: on mathematics, analysis, computer vision technology itself. And practically exercise too.

Where to start, how to continue? By the way, I study C ++ according to Prat's book, I will comprehend Python with Lutz.

And finally: will this direction be relevant in the coming years until the end of the days set aside for meOr are all my desires in vain?

Thank you for reading this long and tedious story)

Answer 1

Knowledge of algorithms is only needed to solve problems that have already been solved before.
For example, there are many algorithms for the same sorting, and among them there is no universal one that would work the fastest in all cases (usually it depends on the volume and mix of data).
Of course, you can superficially get acquainted with the algorithms, but in practice it is worth using ready-made solutions if there are no special requirements.
Algorithms in 3D graphics can be found both in theory and with OpenGL / Direct3D to play around (and to not go far, you can WebGL).
By comp. vision, IMHO, it's better to just start with lectures
On C ++ I can advise
As for long-lived languages, books 4 years ago are quite relevant (you can read about all the features of the new specifications in new books from the authors of the language, which, in my opinion, are more boring).
In other areas, there is nothing special to add: in mathematics, in general, the books of the early 2000s are quite relevant)))
But the relevance of these areas will undoubtedly be in one of the first places. After all, the computing power of computers is increasing, and more and more tasks require automation. Autopilots on electric cars, "smart" video surveillance cameras, etc. - all this is already real. But somehow in Russia this does not add up

Answer 2

you are incredibly lucky with practice,
well, Python + OpenCV, as it were, not in Russian, there are even
academic books, including on the root tracker, and conferences are held

Answer 3

ShAD: https://yandexdataschool.ru/edu-process/courses
General lectures on theory (in pictures): here

Answer 4

Computer Vision section in MatLab.