OCR and text to speech

Hello is there a way to use the camera to do OCR? Also is there a way to do text to speech using some library?

My sister works with disabled people who can't read and asked me to look into it.

I haven't found much in existence right off the shelf. The idea would be to create one app that does both as seamlessly as possible so that with as few steps as possible one could hear a letter being read to them ... for instance.

Sorry if this is quite general.

You could do OCR with a barcode scanner type device - but even then OCR is very 'hit and miss'.

Agraham already did a Text to Speech device library here : http://www.b4x.com/forum/additional-libraries/1250-speech-library-device.html#post6582

What I'd recommend if you're trying to do this is have the text you wanted them to be able to hear embedded in like a QR code or datamatrix barcode, then instead of the issues you'd have with OCR, they'd scan one thing and hear the embedded text.

With OCR - they'd need to pass the scanner/camera over each letter ... it becomes very problematic.

You could use Agraham's barcode library for the PC - http://www.b4x.com/forum/additional-libraries/4062-barcode-generation-library.html#post23517

1) develop a PC app to take a text field and convert it to a barcode.
2) print the barcode
3) have a scanner/camera device read the barcode, then take the received data and use the text to speech library.

The most expensive part of the whole thing would be the barcode readers - the cheapest ones to do this sort of task that could use B4P that my company sells range from like $500 - $750 (but these are industrial devices designed to take a beating...)...

Agraham did write a barcode reading library for devices with a camera with the quickmark application installed - so you could try that too if you had some cheaperish / used devices...

http://www.b4x.com/forum/additional-libraries/4044-quickmark-1d-2d-barcode-reader-library.html

Tex8503 said:

...
1) develop a PC app to take a text field and convert it to a barcode.
...

Click to expand...

Once you have done that you have lost the ability to take the device to the letter rather than vice versa (which I don't think is very practical).

And you might just as well let the PC do the OCR (plenty of apps come with scanners to do that) and the text-to-speech.

I have a pocketable OCR hand-scanner "Pocket Reader" by Siemens bought as a toy (meaning quite cheap). But
a) it no longer works
b) it never did the text-to-speech.

Other devices may be available.

Mike (long-term volunteer with UK RNIB).

mjcoon said:

Once you have done that you have lost the ability to take the device to the letter rather than vice versa (which I don't think is very practical).

Click to expand...

I was thinking along the lines that the letters would be given to the users. The idea of having the PC generate the datamatrix code from the text was prep work before it'd be given to a user. IE : here's the letter (barcode) and the way to 'read' it (IE the device)