IBM Adaptive OCR Engine
Produced by: IBM Israel – Science and Technology Ltd
Scenario
Use this tool when OCR quality is between 50% to 97%. The tool accepts image files with / without extended ALTO transcription files. At the moment, we have a short character session for any new font being processed. The output of the system is ALTO transcription.
Abstract
IBM Adaptive OCR is a comprehensive software system which improves the recognition of historical texts significantly by applying adaptivity as one of the main features to the text recognition process.
It integrates several other tools, such as the image enhancement toolkit, the ABBYY FineReader Engine, the post correction tool and the lexical resources developed during the IMPACT project.
Publications
- Kluzner, V., A. Tzadok, D. Chevion and E. Walach. “Hybrid Approach to Adaptive OCR for Historical Books.” ICDAR2011, 18-21 September, Beijing, China.
- Tzadok, A.
Historic European Texts to be Digitized on a Massive Scale
Availability
The tool is available under commercial licence. For further information on licencing, please contact IBM-Israel IMPACT Group.