Dear all,

Some of you already know that we've been exploring Google's Tesseract and OCRopus projects to see how we can leverage these in building a Bangla OCR. We just released the version 0.5 based on Tesseract's recognition engine for both Windows and Linux. 

This is really an alpha release aimed to get feedback from those who are familiar with the Bangla script. We hope to release a 1.0 version sometime early next year. 

Windows: http://crblpocr.blogspot.com/2008/11/banglaocr-v-05-released.html
Linux: http://crblpocr.blogspot.com/2008/11/banglaocr-v-05-for-linux-users-released.html

You can send feedback via the blog.

The source code, coming soon after a bit of cleanup is done, is released under GNU GPL v2, which is compatible with Tesseract's use of Apache License, version 2.0.

The effort to use OCRopus' recognition engine "bpnet" will come later at some point.

Regards,

Mumit