Let the OCR do the talking for you
I've covered various interesting and non-conventional uses of OCR. I would like to talk about a new one OCR to Speech. The blind community is familiar with technology and it assists them in their everyday lives. The key to OCR to speech is simplicity. When the concept was first developed it required some very elaborate combination of software and hardware, now it's possible to take the latest and greatest OCR technology and make it talk for you with a simple configuration.
It requires a document scanner with a easy physical button interface and programmed to scan an image at 300 DPI to a folder on a machine. Traditional documents work very well for OCR to speech, documents that have a lot of graphics and un-traditional formats may be more challenging. It's important that the technology is able to omit garbage. To do this the OCR process should be driven by a dictionary. The words recognized must be in this dictionary or they will not show up in the final results. The reason for this is a lot of time can be wasted if bad recognition results are spoken.
Once the OCR engine has done it's job of accurately and automatically converting and image to text, the ASCII text results from OCR will be saved into a directory. Now it's time to automatically put the text to speech. There are many text to speech applications out there, some free, some for pay. The goal is to find one that also reads results from a directory and automatically speaks the text over computer speakers.
It can be that easy! Some users of such technologies spend more time trying to find an acceptable digital voice then really configuring the solution. I assure you the packages exist and when configured correctly is very accurate. One scanner, One OCR application hot folder driven, and one text to speech application also hot folder driven will give a robust OCR to speech solution that can be setup in minuets.
It requires a document scanner with a easy physical button interface and programmed to scan an image at 300 DPI to a folder on a machine. Traditional documents work very well for OCR to speech, documents that have a lot of graphics and un-traditional formats may be more challenging. It's important that the technology is able to omit garbage. To do this the OCR process should be driven by a dictionary. The words recognized must be in this dictionary or they will not show up in the final results. The reason for this is a lot of time can be wasted if bad recognition results are spoken.
Once the OCR engine has done it's job of accurately and automatically converting and image to text, the ASCII text results from OCR will be saved into a directory. Now it's time to automatically put the text to speech. There are many text to speech applications out there, some free, some for pay. The goal is to find one that also reads results from a directory and automatically speaks the text over computer speakers.
It can be that easy! Some users of such technologies spend more time trying to find an acceptable digital voice then really configuring the solution. I assure you the packages exist and when configured correctly is very accurate. One scanner, One OCR application hot folder driven, and one text to speech application also hot folder driven will give a robust OCR to speech solution that can be setup in minuets.
Labels: blind, OCR, ocr to speech, text to speech
0 Comments:
Post a Comment
<< Home