opencv - search for known logotype(s) in an image -


I would love to write a program that will take a scanned invoice (original A4 paper, scanned as JPEG file Has been in a PDF), ~ 4000 pixels wide) and look for the logo. If a logotype is found, the invoice file (PDF) will be tagged with the tag tagged with the logo found in the invoice.

I hope 20 or so logotype look and about 2500 invoices, yes, a pain to manually).

My thoughts are to draw towards OpenCV because I know that the sequel has been used behind the curtain. I only want to see the people in some areas, that is, the people only need to move the top left corner of each invoice , Logo b, top right etc. Except JPG for monochrome from high contrast, I think that will help even more? To use keypoints (corners, blobs, etc.), "20 or so logos" is a good number and it's a descriptor (SIFT, Surf, throw etc.) to find the nearest-neighbor-route.

1 train
  • Create a training set of people (take it from your documents)
  • Calculate the set of key points and this details < / li>

    search 2
    • Image equation and noise filtering
    • Find the main point and it is the descriptor
    • The most Good matching narrator (looks nearby neighbor) in you
    • Matching keypoints is not a whole logo but just a casual point < To get the habit to find the homography set training. All the steps are implemented in OpenCV but you will need some time to play with the parameter to keep the best solution. Anyway you have a very low levels of logo distortion, so you will have low levels of "true positive" result and "false positives".

Comments

Popular posts from this blog

jasper reports - How to center align barcode using jasperreports and barcode4j -

Verilog Error: output or inout port "Q" must be connected to a structural net expression -

c# - ASP.NET MVC - Attaching an entity of type 'MODELNAME' failed because another entity of the same type already has the same primary key value -