XF 1.5 OCR usage for converting photos to text fields

[Ludachris]

I have an idea for a plugin that I am trying to spec out. It would involve allowing members to take photos of a standard document type that is very common in my niche and then upload it to the forum, which would then be converted to text using OCR so that the text could then be compared with existing fields in the database. I realize this could be a very involved project. I'd like to get some opinions from developers on how to attack this.

The first question would be, is it even possible to use OCR for something like this in Xenforo? I don't know if OCR technology can be leveraged for something like this. Is there a better way to achieve this?

 

[Jeremy P]

I can't see why it wouldn't be possible, but it'd probably be a little tricky. If you can install tesseract on the server, you can use a PHP binding to grab the OCR text from the upload and go from there.

 

[AndyB]

I have an idea for a plugin that I am trying to spec out.

OCR is not very accurate, so if accuracy is important to you, this technology may not be suitable for what you want to do.

 

[Ludachris]

OCR is not very accurate, so if accuracy is important to you, this technology may not be suitable for what you want to do.

Any other suggestions for processing photos Andy? This is mainly for users taking photos of a specific type of document with their phones, extracting the text to try and match to specified database fields, allow the user to preview what data was extracted, and save/update those fields.

 

[Digital Doctor]

Look into OCR as a service.
Tesseract web service is an option on github

 

[Ludachris]

Look into OCR as a service.
Tesseract web service is an option on github

I'm looking into Tesseract actually. We'll see if I can get this to work.