0

I have a huge data base of images. I want to do a content based image retrieval from it. Now the existing image processing algorithms would take time to search and match among large number of images in the database.

So I first want to retrieve a small subset of images from the database. I can't use an existing image processing algorithm to do that because it takes time.

Now I want to know if there are some basic algorithms that would just tell me if the image could be a match in a very very small amount of time, so that I can just have a small subset of say a thousand images that I would retrieve from the database. I was probably thinking of writing a small histogram based algorithm but it would save a lot of time if something like that was available.

marc_s
  • 732,580
  • 175
  • 1,330
  • 1,459
user3188252
  • 1
  • 1
  • Perceptual hash? Average colour? Per channel average colour? How about getting started on pre-computing some info for all your existing images? Do you expect to match a GIF to a JPEG if the content is similar? Think about rotations, resized images too. – Mark Setchell Oct 08 '14 at 22:14
  • Any pointers to a good library which can help me quickly get started on this ? Yes i expect to match based on content so it should work for different types of images. – user3188252 Oct 08 '14 at 22:19

0 Answers0