Jump to content

Recommended Posts

Posted

Hello all.

I'm trying to do something simple, but I can't manage to find where to start since I can have multiple possibilities and not sure what is the best one :

  1. Taking screen capture (I can do that)
  2. On this capture, I'll look for 10 particular regions that will give me 10 images of 40px * 40px (I can do that)
  3. Then what I want is for each of these regions, find on my library of ~140 images which one is the closest one to my region (= the match)

 

What do you think is the best approach ? Ican easily do step 1 and 2, however for the third step :

  • I can't use memcmp because they won't be exactly the same (original library can differ a very little bit from what I'll get from my screen capture)
  • There is no text, only drawing
  • Should I do a pixelSum of what I get on screen and do the same via GDIPlus for each of my images in the library?

 

Thanks a lot,

Tim

 

Posted (edited)

Tried this ;

And this :

Without a lot of success, bot scripts fail for different reason and are quite old.

Edited by timmalos
Posted
1 hour ago, timmalos said:

I'm trying to do something simple

Errm, no, not at all. Image recognition and interpretation is computationally very hard, otherwise companies wouldn't be spending hundreds of millions of dollars on developing better algorithms (self-learning, AI, PCA) to do something you take for granted such as distinguishing between the faces of your family members, for example.

That said, here are some simple suggestions, or you could use ImageMagick's compare functionality.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...