There has been long waiting for a Bangla OCR (optical character recognition) software. Personally I have been involved in Bangla OCR research in my spare time for a few years. At last I have the feeling of having something done.

This version of Bangla OCR is merely an alpha. It has a very high recognition speed, but the accuracy is not something to boast. It surely has bugs and weakness, and I am working on these.

Download the Bangla OCR here. 

Good things about Apona Pathak Bangla OCR-

• Fast recognition speed.
• Direct scanning from scanner.
• Support for popular image formats.
• Automatically correctc document skew resulted from casual scanning.
  
• Can send output to MS Word automatically.
• Can produce UNICODE texts.
• and it's free.

Limitations of Pathak-

• Only Sutonny and the like font faces are supported.
• Only 10 to 14 point sizes are recognized.
• No support for italic texts.
• Can not recognize formatting.
• No support for multi-column texts.
• Can not handle document with picture and tables.
• Can not recognize English texts in Bangla documents.

To do list (according to priority)-

• Increasing recognition rate.
• Finding memory leaks and other bugs.
• Multiple font support.
• English texts support.
• Photographs and table support.