2
$\begingroup$

Excuse my ignorance...

Given a digital image of a label I'd like to

  • Requirement: Match it to the closest pre-existing image in a datastore
  • Optional: OCR it, if possible (only necessary if step 1 fails)

I'm a programmer but not an engineer. I have no idea where to begin or where to look into solving this problem. I don't even know the right terminology; "Image Processing" is best I could come up with.

Starting here, on DSP (beta); apologies in advance to the forum police if this type of inquiry is not considered to be 'in scope'. Before you vote me off the 'island' hopefully someone can comment with a better site/forum for my education.

$\endgroup$
5
$\begingroup$

First, you will need an image processing algorithm to detect local features of the digital images you have. 1 and 2 provide good reviews of such algorithms.:

After extracting the local features, a matching process between the extracted features and your set of labels is used to link the labels with acquired images. The second references goes over few matching techniques and performance metrics.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.