OCR system for the printed Indian Scripts converts scanned document
images into editable Unicode text format. The system has been
developed for Hindi, Bangla, Tamil, Gurumukhi, Malayalam and Odia
Script. OCR system bundled with pre-processing and post-processing
algorithms to provide end-to-end solution. Solution is having a hybrid
approach, designed to work with the platform and technology
Product for the end-to-end solution available to work on Windows (Windows 7) and Linux (Ubunto, Fedora) based operating system. The web based solution is also developed for the OCR services.
Main Uses and Domain
- Conversion of Old archived document images into text for information retrieval and indexing for Heritage Computing.
- Generation of Unicode text from text books images and to make the content ready for the other processing like TTS, Machine Translation etc.
- Data Entry automation
- ebook generation
Features and Technical Specifications
- Basic image enhancement and editing tools (cropping, rotation, zoom in/zoom out, orientation, binarization, noise removal etc.
- Works on Grey level and black 'n' white images
- Support for BMP and PNG formats.
- Running individual modules successively for obtaining final OCR output.
- End-to-End OCR
- More than 90% accuracy at character level
- Best performance when scanned at 300 dpi
- Text editing tool coupled with dictionary
- Output in Unicode supported format
Platform required(if any)
- Desktop Version:Windows 7 and above
- Desktop version for Linux(Ubuntu, Fedora) Operating System
- Web Based
Contact Details for Techno Commercial Information
- Mr. Tushar Patnaik
Principal Technical Officer,SNLP Lab, CDAC-Noida
Email ID: tusharpatnaik[at]cdac[dot]in
- Prof. Santanu Chaudhury
Director, CEERI PILANI
Email ID: director[at]ceeri[dot]ernet[dot]in