RWTH OCR - The RWTH Aachen University Optical Character Recognition System
RWTH OCR is a software package containing an OCR decoder together with tools for the development of visual models, based on RWTH ASR. It has been developed by the Human Language Technology and Pattern Recognition Group at the RWTH Aachen University. OCR systems developed using this framework have been applied successfully in several international research projects and corresponding evaluations.
RWTH OCR consists of several libraries and tools written in C++. Currently, only Linux (x86 and x86-64) platforms are supported.
Features
- decoder for OCR with large vocabulary
- feature extraction
- a flexible framework for data processing:
Flow
- easy implementation of new features as well as easy integration of external features using Flow networks
- visual modeling
- Gaussian mixture distributions for HMM emission probabilities
- language modeling
- support for language models in ARPA format
- adaptation
- Constrained MLLR (CMLLR, "feature space MLLR")
- Unsupervised maximum likelihood linear regression mean adaptation (MLLR)
- input / output formats
- nearly all input and output data is in easily process-able XML formats
- converter tools for the generation of NIST file formats are included
Documentation
The development of RWTH OCR is ongoing. A Manual is available in the Wiki. Access to the wiki requires registration.
Publications about the theoretical foundations and methods used can be found in the publications page. RWTH ASR is described in detail in Rybach et al. The RWTH Aachen University Open Source Speech Recognition System. Interspeech 2009.
Please post questions in the support forum.
Installation
RWTH OCR is an add-on package for RWTH ASR.
RWTH OCR is available only in source form. See the included README.RWTH-OCR for build instructions.
A set of installed tools and libraries is required (Debian package name given in brackets):
- RWTH ASR 0.6.1
- RWTH ASR dependencies (see README)
- PNG library (libpng12-dev)
- JPEG library (libjpeg62-dev)
- Netpbm (libnetpbm10-dev)
After downloading and extracting RWTH ASR, apply the RWTH OCR patch:
cd rwth-asr-0.6.1
gzip -dc rwth-ocr-0.1.patch-for-0.6.1.gz | patch -p1
Terms of Use
RWTH OCR is free software; it can be redistributed and/or modified under the terms of the RWTH OCR License. This license includes free usage for non-commercial purposes as long as any changes made to the original software are published under the terms of the same license. Other licenses can be requested.
Download
To download the software, you have to accept the license terms. Please fill out
Related Work
-
P. Dreuw, D. Rybach, G. Heigold, and H. Ney. RWTH OCR: A Large Vocabulary Optical Character Recognition System for Arabic Scripts. In Volker Märgner, and Haikal El Abed: Guide to OCR for Arabic Scripts Chp. Part II: Recognition, pages 215-254, Springer, London, UK, July 2012.
ISBN 978-1-4471-4071-9.
©
-
P. Dreuw. The RWTH OCR System For Handwritten and Machine Printed Text Recognition. Google Tech Talk, Mountain View, CA, USA, January 2011.
-
G. Heigold, P. Dreuw, S. Hahn, R. Schlüter, and H. Ney. Margin-Based Discriminative Training for String Recognition. IEEE Journal of Selected Topics in Signal Processing - Statistical Learning Methods for Speech and Language Processing, volume 4, number 6, pages 917-925, Aachen, Germany, December 2010.
-
P. Dreuw, and H. Ney. The RWTH-OCR Handwriting Recognition System for Arabic Handwriting. DAAD Workshop III - On the Way to the Information Society, Sousse, Tunisia, March 2010.
Invited Talk.
-
P. Dreuw, G. Heigold, and H. Ney. Confidence-Based Discriminative Training for Model Adaptation in Offline Arabic Handwriting Recognition. In International Conference on Document Analysis and Recognition (ICDAR), pages 596-600, Barcelona, Spain, July 2009.
©
-
P. Dreuw, D. Rybach, C. Gollan, and H. Ney. Writer Adaptive Training and Writing Variant Model Refinement for Offline Arabic Handwriting Recognition. In International Conference on Document Analysis and Recognition (ICDAR), pages 21-25, Barcelona, Spain, July 2009.
©
-
P. Dreuw, S. Jonas, and H. Ney. Advances in Arabic and Latin Handwriting Recognition using the RWTH-OCR System. Traitement de la parole, du langage et des documents multimedias (Seminaire DGA), Paris, France, July 2009.
-
S. Jonas. Improved Modeling in Handwriting Recognition. Master Thesis, Aachen, Germany, June 2009.
-
P. Dreuw, S. Jonas, and H. Ney. White-Space Models for Offline Arabic Handwriting Recognition. In International Conference on Pattern Recognition (ICPR), pages 1-4, Tampa, FL, USA, December 2008.