Comments on: OCRFeeder 0.7.7 released http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/ Free Software and travelling. Fri, 10 May 2013 13:05:04 +0000 hourly 1 http://wordpress.org/?v=3.5.1 By: Joaquim Rocha http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/comment-page-1/#comment-6573 Joaquim Rocha Wed, 25 Jan 2012 19:43:27 +0000 http://www.joaquimrocha.com/?p=1214#comment-6573 Don’t worry, happens all the time :)

]]> By: Russell Hedger http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/comment-page-1/#comment-6572 Russell Hedger Wed, 25 Jan 2012 19:42:08 +0000 http://www.joaquimrocha.com/?p=1214#comment-6572 I can’t even spell it correctly, Joaquim.

]]> By: Russell Hedger http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/comment-page-1/#comment-6571 Russell Hedger Wed, 25 Jan 2012 19:40:19 +0000 http://www.joaquimrocha.com/?p=1214#comment-6571 Ha, sorry: Joachim, not Jürgen!

]]> By: Joaquim Rocha http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/comment-page-1/#comment-6570 Joaquim Rocha Wed, 25 Jan 2012 17:27:24 +0000 http://www.joaquimrocha.com/?p=1214#comment-6570 I can add the plain text exportation feature to the CLI for the next version.

BTW, who is Jürgen? :)

]]> By: Russell Hedger http://www.joaquimrocha.com/2011/12/10/ocrfeeder-0-7-7-released/comment-page-1/#comment-6568 Russell Hedger Wed, 25 Jan 2012 16:21:55 +0000 http://www.joaquimrocha.com/?p=1214#comment-6568 Hello Jürgen,

I am using OCRFeeder from the repositories on Ubuntu 11.10, and it works quite well with Cuneiform as the OCR engine. However, when using the latest stable version of tesseract-ocr 3.01 (which I have found to be more accurate at OCR), character recognition does not work properly in OCRFeeder. Perhaps it is related to this issue:

http://code.google.com/p/tesseract-ocr/issues/detail?id=580

Also, I have been experimenting with using the OCRFeeder cli to feed scanned print into a speech engine as an aid for someone with a visual impairment. The gui does a pretty good job at recognising text boxes, performing ocr and exporting as plaintext. I have found that tesseract alone reads print quite well but gets confused by artifacts such as staples, edges and line graphics. OCRFeeder is much better at dealing with this issue, and passing the correct image parts to tesseract to be recognised. I notice that the cli only exports as html and odt, and it would be very helpful if any future version had the ability to output in plaintext, which could then be passed straight to a speech engine.

]]>