Video recordings processing

Python Help
1

Hi all,

I’d like to write a small software to process different tv recordings I did over time, to identify overlaps between videos files and to reorganize them accordingly.
Files are in mkv format.

How would you approach this problem?

Shall I use libraries like opencv?

Thanks for your answers

Marco

2

Your question is a little unclear - what do you mean by “organizing them accordingly”? According to overlaps? If so, what do you mean by that?

That aside, personally I always try to look for an algorithmic approach before touching machine learning, which has lead to exactly 0 times I’ve needed machine learning/artificial neural nets. The current state of the art, while impressive, is still not trivial to pick up if you have no prior education or experience, and wont give perfectly reliable results. I would tend to only use machine learning when I don’t care if the results are perfect or if there is no clear algorithmic approach. For example, sorting & categorizing an overwhelming number of images, where even a human is likely to make as many mistakes. I also can’t help but think a consumer-grade computer is going to be very very slow at analyzing video content.

Here are a couple of questions you should probably ask yourself:

  • Are these overlaps consistent in time? If so, you don’t need the computer to “look at it” for you, unless you’re not sure which files go with each other.
  • How many files are we talking about? Are you sure it’s worth the time/effort for a presumably disposable program? I have found that a lot of times when I want to write a program, it’s actually easier and faster to do the work manually.

I realize that’s not terribly helpful, but if you clarify your question a little more, perhaps I or someone more experienced would be able to answer.

Cheers.