0

I've the following issue:

  1. starting from an image (example: a picture of my food pantry) I have to split in sub-images according to different objects contained in it.
  2. after having isolated each object (and got its image) preprocess the image with split() function dividing channels
  3. detect text contained in each image (example "peanuts", on the peanuts pack) and read text contained in it (using tesseract libs)
  4. detect symbols in each image (example coca-cola logo, on a coca-cola bottle) using a SURF library...

Explained my purpose, the question is: - how can I perform activity 1?

Avi
  • 21,182
  • 26
  • 82
  • 121
user2875719
  • 1
  • Have you tried pattern recognition and subsequent Region Of Interest (`ROI`) around each object (pattern that has been recognised)? I was working on something roughly similar and did it like this. But my problem was easy because the pictures were always taken from the same distance so a constant size of `ROI` was just fine (objects had the same size). – Aleksander Lidtke Oct 13 '13 at 09:21
  • hi, thanks for your answer, I've not tried it so I'll check this way. For the distance from which photo is taken this is not a problem (I mean that I'll keep it constant, if it's required..it's a constrain I can accept). I have one doubt however: does pattern recognition work if I have surmounted objects? (one on the other, or one partially covering an other)? – user2875719 Oct 13 '13 at 10:36
  • well you don't need to keep it constant if you figure out a way to scale the region of interest's size so it only leaves the object you're interested in (could e.g. find the corners with pattern recognition). As for the surmounted objects it should if it can detect the features of a given object (could have a set for each object so something distinct is always visible). – Aleksander Lidtke Oct 13 '13 at 11:29

0 Answers0