Arabic OCR
من ويكي عربآيز
محتويات
Optical Character Recognition
OCR is the ability to scan a document (or grab a PDF file) and run an OCR program on it and it will generate, based on optical recognition and approximation, an editable text file. For an idea about OCR see http://www.students.cs.uu.nl/people/mjkammer/Work/intro_2_OCR.html
Current Status of Arabic OCR software
I (MuhammadAlkarouri) know of no actually working Arabic OCR software that is open source. Any additions are certainly welcome.
Resources
- List of Linux OCR applications: http://www.linux-ocr.ekitap.gen.tr/
Arabic OCR Links
- Automatic Recognition Using Zernike Moments As A Feature Extractor (Paper) http://www.ici.ro/ici/revista/sic2001_3/art4.html
- Graph Based Segmentation .. (Paper) http://ceri.kacst.edu.sa/webpage/software_a_3.htm
- Structural Features Of Cursive Arabic Scripts (Paper) http://www.bmva.ac.uk/bmvc/1999/papers/42.pdf
- Multilingual Machine Printed OCR (Paper) http://portal.acm.org/citation.cfm?id=505744&dl=ACM&coll=GUIDE
- Test of two Arabic OCR programs http://www.hf.uib.no/smi/ksv/arabocr.html
- Performance Evaluation of two Arabic OCR products http://www.ai.mit.edu/~gremio/publications/Kanungo-etal-AIPR98.pdf
Other Links
- Software from SA http://ceri.kacst.edu.sa/webpage/software_a_3.htm
- How to encode image produced by a recognition system (mailing thread) http://lists.arabeyes.org/archives/general/2002/March/msg00001.html
- Rapidly Retargetable Translingual Detection http://tides.umiacs.umd.edu/description.html
- Sibawayhi Project http://www.hf.uio.no/east/sibawayhi/HomePage/