OPTICAL CHARACTER RECOGNITION (CONVERTING RESTAURANT RECEIPTS INTO TEXT)

Dear all,

I am working on converting image of restaurant receipts into readable text using OCR. I have already completed sort of pre processing steps for OCR with help of http://basic-eng.blogspot.com/2006/04/character-recognition-example.html

Just to brief on what is done so far is that the image is first converted to grayscale, followed by converting into binary image using threshold technique then the binary image has gone through connectivity test in order to check for the maximum connected component which is the box of the form. The final result is as shown in attachement

Now the problem is after locating the box, how can i crop individual character into different subimages that will be the raw data for classification technique.

Any help on this is much appreciated :)

Thanks :)

Answer 1

[Dima Lisin]

You can use imcrop for cropping.

But better yet, try the ocr() function in the Computer Vision System Toolbox.

[Isra Abbasi]

Thanks for the suggestion. But I don't have Computer Vision System toolbox in my version of matlab. Do u know, from where can i get one ?

Thanks again

[Dima Lisin]

Hi Isra,

You can buy it from Mathworks. Also, you would need a recent version of MATLAB. ocr() was added in R2014a.