CODEWITHSUNDEEP
pythonocr
Osmond Bishop
Osmond Bishop

Reputation: 8180

Creating a table from an image in python

I've never used OCR before and I was wondering if anyone can recommend a decent OCR for python that would be able to convert the following image into a 2D array?

enter image description here

Upvotes: 0

Views: 539

Answers (1)

Ilya Evdokimov
Ilya Evdokimov

Reputation: 1394

OCR stands for Optical CHARACTER Recognition. Any generic OCR, even paid high-quality commercial product, will be very unreliable on this image. I tried one - it produces some dots, and keeps most other areas as pictures due to lack of characters.

To process this reliably, especially if the volume is high, I would write an image analysis routine (if images are consistent) or use some specialized customizable tool. To be specific, I would use ABBYY FlexiCapture (because I have access to it) and create a quick template to detect a matrix of 'objects' (black blobs) in table-like pattern. 10 minutes of work at most.

Upvotes: 2

Related Questions