Arabic open source OCR

[David Gelbart <zizazze at yahoo dot ca>]

I just read the September 2007 discussion about Arabic OCR using HTK.  Sounds great.

Part of the discussion concerned the distribution restrictions on HTK.  As was mentioned in September, those restrictions are an obstacle to creating a Debian package that is capable of a fully automated install without human intervention.    I'm not sure if this issue is important at this stage, but anyway here are some thoughts:

1) It might be possible to replace HTK with another recognition engine.  VoxForge  (www.voxforge.org) lists several open source recognition engines.

2) If OCR users will not normally be creating their own models, it might be possible to use HTK to train the models, but package another recognition engine for recognition (such as Julius, which can read HTK models).    

Regards,
David

---

Be smarter than spam. See how smart SpamGuard is at giving junk email the boot with the All-new Yahoo! Mail