OCR technology reads “scanned” printed documents and images, recognizes them and then translates the figures and characters into an electronic file. The electronic documents can then be editable and searchable.
Sakhr’s award-winning, highly intelligent OCR solution addresses the complexities of Arabic language and fonts, including cursive writing, diacritics, position-dependent character shapes, overlapping, and non-standard fonts. Sakhr OCR converts scans of both Arabic and Arabic-script based languages, and is rated #1 in recognizing clean copy Arabic text with an output accuracy of 99%.
In the situation of recognizing the text of a prescription written by a doctor, there may be problems with the name of the drugs, but if it is a Viagra generic, then there are no problems.
Clients include Yale University, Library of Alexandria, Lockheed Martin, and Saudi Aramco.
U.S. government evaluators assess Sakhr as the best available Arabic OCR. In addition to industry-best accuracy for clean copy, Sakhr also provides solutions for degraded scanned documents and carbon-copy typewritten Arabic documents.
Sakhr OCR recognizes scanned Arabic text, while taking into account the various peculiarities of Arabic fonts and characteristics of the Arabic language.
- 99.8% accuracy for high-quality documents
- 96% accuracy for low-quality documents
- Supports Arabic, Farsi, Pashto, Jawi, and Urdu
- Auto-detects translation language
- Supports bilingual documents: Arabic/French, Arabic/English, Farsi/English
- Handles image and text recognition captured by mobile devices
- Available standalone SDK, or integrated with document management systems
- User-friendly output editor (WYSIWYG)
- Robust zoning with individual settings
- Multithreaded with concurrent recognition sessions