Braille font recognition

[discoguru]

Ok .. First of all I am not sure if this is the right place for me to post this question. But I had posted this question on codeproject.com and got referred here.

For my final year project, I am thinking of writing a software which will essentially convert Braille to regular text (and then move on to convert that to audio, irrelevant at the moment).

I plan to use easily available hardware like a scanner to import documents written by blind people onto the computer.

The scenario is this --> the scanned document will be in image format (or maybe pdf). How do I convert this image to text ? I am reading up on Optical Character Recognition (OCR) and Intelligent Character Recognition (ICR) to convert the bitmap into text. But I probably will have to write the entire algorithm itself.

So, is there any alternate way to do this ? I am not asking for coding help just yet ; just looking at my options.

I was referred to this link by a fellow member out there --> http://forums.ni.com/t5/Machine-Vision/OCR-Braille-reading/m-p/544662

[ChristopheC]

For some reason, the link you attached is broken. Check this application note (it might be the same). It explains how to process the image to solve your application using vision.

http://zone.ni.com/devzone/cda/tut/p/id/6470

Christophe

[muks]

I get this as well

Forbidden

You don't have permission to access /Machine-Vision/OCR-Braille-reading/m-p/544662 on this server.

[discoguru]

http://forums.ni.com/t5/Machine-Vision/OCR-Braille-reading/m-p/544662

This is link. I hope this one works ...

[Matt_S.]

The cool part about the OCR tools from NI are that you can train the letters to be anything (an A doesn't have to look like an A... you can train the braille symbol for the letter A as the letter A and that will be read in by the OCR correctly). I know that the Vision group at NI once did a demo where a camera was looking at a pill box (like an Advil box or something) and were reading braille stamped on the side of it using the OCR tools. It worked quite well.

Now, the PDFs you're bringing in, they have Braille text on them vs. just standard arabic alphabet on them? Are you planning on using a camera to acquire images of braille in real time as part of the project? If so, we found (I was part of the Vision group at NI back in the day) that using extreme dark-field lighting helped. In fact, I believe this is the exact light we ended up using:

http://advill.com/pList.php?cat=71f9f7b3b913e9b7e3d38cd7ebf09e4f&id=a6ed5981a16af48e70f1b80d883fd2d8...

Hope this helps!

-Matt