Printed Ethiopic Script Recognition by Using LSTM Networks

Bidirectional Long Short-Term Memory (LSTM) networks have brought tremendous results on many machine learning tasks including handwritten and machine printed character recognition systems. The Ethiopic script uses a large number of characters in the writing and existence of visually similar character, which results in a challenge for OCR development. In this paper, we present application of bidirectional LSTM neural networks to recognize machine printed Ethiopic scripts. To train and test the model, we collect text files from different source written in Amharic, Ge’ ez and Tigrigna language and generate 96,000 artificial text line images by applying different degradation techniques. Additionally, to test the model with real scanned documents, we use real 12 page scanned images from Tsenat book. Without using any language modeling and any other post-processing, LSTM networks attain an average character error rate of 2.12%, and this indicates the proposed network achieves a promising result.

[1]  Yaregal Assabie,et al.  Multifont size-resilient recognition system for Ethiopic script , 2007, International Journal of Document Analysis and Recognition (IJDAR).

[2]  Jürgen Schmidhuber,et al.  LSTM recurrent networks learn simple context-free and context-sensitive languages , 2001, IEEE Trans. Neural Networks.

[3]  Berrin A. Yanikoglu,et al.  Off-line cursive handwriting recognition using neural networks , 1993, Defense, Security, and Sensing.

[4]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[5]  Saad Bin Ahmed,et al.  Offline Printed Urdu Nastaleeq Script Recognition with Bidirectional LSTM Networks , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[6]  Jürgen Schmidhuber,et al.  Finding temporal structure in music: blues improvisation with LSTM recurrent networks , 2002, Proceedings of the 12th IEEE Workshop on Neural Networks for Signal Processing.

[7]  Bi Liu,et al.  A Normalized Levenshtein Distance Metric , 2007, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Richard M. Schwartz,et al.  A Script-Independent Methodology For Optical Character Recognition , 1998, Pattern Recognit..

[9]  Jan Nordin,et al.  Offline OCR System for Machine-Printed Turkish Using Template Matching , 2011 .

[10]  Li Sun,et al.  Deep LSTM Networks for Online Chinese Handwriting Recognition , 2016, 2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR).

[11]  Yaregal Assabie,et al.  A neural network approach for multifont and size-independent recognition of ethiopic characters , 2007 .

[12]  Thomas M. Breuel,et al.  High-Performance OCR for Printed English and Fraktur Using LSTM Networks , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[13]  Alex Graves,et al.  Connectionist Temporal Classification , 2012 .

[14]  Didier Stricker,et al.  A comparison of 1D and 2D LSTM architectures for the recognition of handwritten Arabic , 2015, Electronic Imaging.

[15]  Narendra S. Chaudhari,et al.  Bidirectional segmented-memory recurrent neural network for protein secondary structure prediction , 2006, Soft Comput..

[16]  Jürgen Schmidhuber,et al.  Learning Nonregular Languages: A Comparison of Simple Recurrent Networks and LSTM , 2002, Neural Computation.

[17]  C. V. Jawahar,et al.  Optical Character Recognition of Amharic Documents , 2007, Afr. J. Inf. Commun. Technol..

[18]  Michel Dhome,et al.  A simple and efficient template matching algorithm , 2001, Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001.