Offline cursive Urdu-Nastaliq script recognition using multidimensional recurrent neural networks

Optical Character Recognition of cursive scripts remains a challenging task due to a large number of character shapes, inter- and intra-word overlaps, context sensitivity and diagonality of text. This paper presents an implicit segmentation based recognition system for Urdu text lines in Nastaliq script. The proposed technique relies on sliding overlapped windows on lines of text and extracting a set of statistical features. The extracted features are fed to a multi-dimensional long short term memory recurrent neural network (MDLSTM RNN) with a connectionist temporal classification (CTC) output layer that labels the character sequences. Experimental study of the proposed technique is carried out on the standard Urdu Printed Text-line Images (UPTI) database which comprises 10,000 text lines in Nastaliq font. Evaluations under different experimental settings realize promising recognition rates with a highest character recognition rate of 96.40%.

[1]  Imran Siddiqi,et al.  Segmentation techniques for recognition of Arabic-like scripts: A comprehensive survey , 2015, Education and Information Technologies.

[2]  Rongrong Ji,et al.  Robust Optical Recognition of Cursive Pashto Script Using Scale, Rotation and Location Invariant Approach , 2015, PloS one.

[3]  Adel M. Alimi,et al.  A New Arabic Printed Text Image Database and Evaluation Protocols , 2009, 2009 10th International Conference on Document Analysis and Recognition.

[4]  Laurence Likforman-Sulem,et al.  New baseline correction algorithm for text-line recognition with bidirectional recurrent neural networks , 2013, J. Electronic Imaging.

[5]  Shuai Li,et al.  Decentralized kinematic control of a class of collaborative redundant manipulators via recurrent neural networks , 2012, Neurocomputing.

[6]  Muhammad Imran Razzak,et al.  Arabic script based language character recognition: Nasta'liq vs Naskh analysis , 2013, 2013 World Congress on Computer and Information Technology (WCCIT).

[7]  Yangming Li,et al.  A class of finite-time dual neural networks for solving quadratic programming problems and its k-winners-take-all application , 2013, Neural Networks.

[8]  Imtiaz Ahmed,et al.  Challenges of Urdu Named Entity Recognition: A Scarce Resourced Languageq , 2014 .

[9]  Jean Camillerapp,et al.  Kalman filter contributions towards document segmentation , 1995, Proceedings of 3rd International Conference on Document Analysis and Recognition.

[10]  Yajin Zhou,et al.  SP-NN: A novel neural network approach for path planning , 2007, 2007 IEEE International Conference on Robotics and Biomimetics (ROBIO).

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

[12]  Luca Maria Gambardella,et al.  Convolutional Neural Network Committees for Handwritten Character Classification , 2011, 2011 International Conference on Document Analysis and Recognition.

[13]  Marc'Aurelio Ranzato,et al.  Large Scale Distributed Deep Networks , 2012, NIPS.

[14]  Shamandeep Singh,et al.  Offline Handwritten Devanagari Numerals Recognition using GLCM Features & Neural Networks , 2014 .

[15]  Farhad Faradji,et al.  A Comprehensive Isolated Farsi/Arabic Character Database for Handwritten OCR Research , 2006 .

[16]  Mohammad Matin,et al.  Urdu character recognition using fourier descriptors for optical networks , 2005, SPIE Optics + Photonics.

[17]  Muhammad Imran Razzak,et al.  Challenges in Baseline Detection of Arabic Script Based Languages , 2014 .

[18]  Jürgen Schmidhuber,et al.  Multidimensional Recurrent Neural Networks , 2007 .

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

[20]  Samee Ullah Khan,et al.  The optical character recognition of Urdu-like cursive scripts , 2014, Pattern Recognit..

[21]  Muhammad Imran Razzak,et al.  Evaluation of cursive and non-cursive scripts using recurrent neural networks , 2015, Neural Computing and Applications.

[22]  Faisal Shafait,et al.  A segmentation-free approach to Arabic and Urdu OCR , 2013, Electronic Imaging.

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

[24]  Gurpreet Singh Lehal.  Ligature Segmentation for Urdu OCR , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[25]  Santanu Chaudhury,et al.  Text recognition using deep BLSTM networks , 2015, 2015 Eighth International Conference on Advances in Pattern Recognition (ICAPR).

[26]  Edouard Geoffrois,et al.  Results of the RIMES Evaluation Campaign for Handwritten Mail Processing , 2009, 2009 10th International Conference on Document Analysis and Recognition.

[27]  Faisal Shafait,et al.  Search Space Reduction for Holistic Ligature Recognition in Urdu Nastalique Script , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[28]  Horst Bunke,et al.  The IAM-database: an English sentence database for offline handwriting recognition , 2002, International Journal on Document Analysis and Recognition.

[29]  Jürgen Schmidhuber,et al.  Unconstrained On-line Handwriting Recognition with Recurrent Neural Networks , 2007, NIPS.

[30]  Tetsuya Ogata,et al.  Handwriting prediction based character recognition using recurrent neural network , 2011, 2011 IEEE International Conference on Systems, Man, and Cybernetics.

[31]  Inam Shamsher,et al.  Urdu compound Character Recognition using feed forward neural networks , 2009, 2009 2nd IEEE International Conference on Computer Science and Information Technology.

[32]  Ehsanollah Kabir,et al.  Introducing a very large dataset of handwritten Farsi digits and a study on their varieties , 2007, Pattern Recognit. Lett..

[33]  Jürgen Schmidhuber,et al.  Multi-dimensional Recurrent Neural Networks , 2007, ICANN.

[34]  Saeed Mozaffari,et al.  ICDAR 2009 Handwritten Farsi/Arabic Character Recognition Competition , 2009, 2009 10th International Conference on Document Analysis and Recognition.

[35]  Mohammad S. Khorsheed,et al.  HMM-based recognition engine using a novel approach for statistical feature extraction , 2014, The 2014 2nd International Conference on Systems and Informatics (ICSAI 2014).

[36]  Shuai Li,et al.  Selective Positive–Negative Feedback Produces the Winner-Take-All Competition in Recurrent Neural Networks , 2013, IEEE Transactions on Neural Networks and Learning Systems.

[37]  R. Manmatha,et al.  Features for word spotting in historical manuscripts , 2003, Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings..

[38]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[39]  Geoffrey Zweig,et al.  Accelerating recurrent neural network training via two stage classes and parallelization , 2013, 2013 IEEE Workshop on Automatic Speech Recognition and Understanding.

[40]  Imran Siddiqi,et al.  An Ocr system for printed Nasta'liq script: A segmentation based approach , 2014, 17th IEEE International Multi Topic Conference 2014.

[41]  Laurence Likforman-Sulem,et al.  The UOB-Télécom ParisTech Arabic handwriting recognition and translation systems for the OpenHaRT 2013 competition , 2013 .

[42]  Hui Zhang,et al.  Decentralized robust attitude tracking control for spacecraft networks under unknown inertia matrices , 2015, Neurocomputing.

[43]  Horst Bunke,et al.  Using a Statistical Language Model to Improve the Performance of an HMM-Based Cursive Handwriting Recognition System , 2001, Int. J. Pattern Recognit. Artif. Intell..

[44]  Akram M. Zeki,et al.  Segmentation techniques for online Arabic handwriting Recognition: A survey , 2010, Proceeding of the 3rd International Conference on Information and Communication Technology for the Moslem World (ICT4M) 2010.

[45]  Muhammad Imran Razzak,et al.  Urdu Nasta’liq text recognition system based on multi-dimensional recurrent neural network and statistical features , 2017, Neural Computing and Applications.

[46]  Ching Y. Suen,et al.  A New Large Urdu Database for Off-Line Handwriting Recognition , 2009, ICIAP.

[47]  Jürgen Schmidhuber,et al.  Framewise phoneme classification with bidirectional LSTM and other neural network architectures , 2005, Neural Networks.

[48]  Marcus Liwicki,et al.  A novel approach to on-line handwriting recognition based on bidirectional long short-term memory networks , 2007 .

[49]  Volker Märgner,et al.  Arabic Handwriting Recognition Competition , 2005, ICDAR.

[50]  M. Kalaiselvi Geetha,et al.  Segregated Handwritten Character Recognition using GLCM Features , 2013 .

[51]  Fadi Biadsy,et al.  Segmentation-Free Online Arabic Handwriting Recognition , 2011, Int. J. Pattern Recognit. Artif. Intell..

[52]  Christopher Kermorvant,et al.  Dropout Improves Recurrent Neural Networks for Handwriting Recognition , 2013, 2014 14th International Conference on Frontiers in Handwriting Recognition.

[53]  Raja Muhammad Asif Zahoor,et al.  Design of stochastic solvers based on genetic algorithms for solving nonlinear equations , 2014, Neural Computing and Applications.

[54]  Shuai Li,et al.  Inverse-Free Extreme Learning Machine With Optimal Information Updating , 2016, IEEE Transactions on Cybernetics.

[55]  Mohammad S. Khorsheed,et al.  Offline recognition of omnifont Arabic text using the HMM ToolKit (HTK) , 2007, Pattern Recognit. Lett..

[56]  Sarmad Hussain,et al.  Adapting Tesseract for Complex Scripts: An Example for Urdu Nastalique , 2014, 2014 11th IAPR International Workshop on Document Analysis Systems.

[57]  Jürgen Schmidhuber,et al.  Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks , 2006, ICML.

[58]  Chee Kheong Siew,et al.  Extreme learning machine: Theory and applications , 2006, Neurocomputing.

[59]  Geoffrey E. Hinton,et al.  Training Recurrent Neural Networks , 2013 .

[60]  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.

[61]  Andrew W. Senior,et al.  Long short-term memory recurrent neural network architectures for large scale acoustic modeling , 2014, INTERSPEECH.

[62]  Muhammad Sher,et al.  HMM and fuzzy logic: A hybrid approach for online Urdu script-based languages' character recognition , 2010, Knowl. Based Syst..

[63]  Fei Meng,et al.  An extended Kalman filter for input estimations in diesel-engine selective catalytic reduction applications , 2016, Neurocomputing.

[64]  Sarmad Hussain,et al.  Segmentation Based Urdu Nastalique OCR , 2013, CIARP.

[65]  Jürgen Schmidhuber,et al.  Offline Handwriting Recognition with Multidimensional Recurrent Neural Networks , 2008, NIPS.

[66]  Mohamed Cheriet,et al.  Feature Design for Offline Arabic Handwriting Recognition: Handcrafted vs Automated? , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[67]  Sarmad Hussain,et al.  Segmentation Free Nastalique Urdu OCR , 2010 .

[68]  Neil W. Bergmann,et al.  An Arabic optical character recognition system using recognition-based segmentation , 2001, Pattern Recognit..

[69]  Kuldip K. Paliwal,et al.  Bidirectional recurrent neural networks , 1997, IEEE Trans. Signal Process..

[70]  Marc-Peter Schambach,et al.  Low resolution Arabic recognition with multidimensional recurrent neural networks , 2013, MOCR '13.

[71]  Venu Govindaraju,et al.  Retrieving Handwriting Styles: A Content Based Approach to Handwritten Document Retrieval , 2010, 2010 12th International Conference on Frontiers in Handwriting Recognition.

[72]  Brijesh Verma,et al.  A novel feature extraction technique for the recognition of segmented handwritten characters , 2003, Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings..

[73]  Shehzad Khalid,et al.  Recognition of Urdu ligatures - a holistic approach , 2015, 2015 13th International Conference on Document Analysis and Recognition (ICDAR).

[74]  Alireza Alaei,et al.  Using Modified Contour Features and SVM Based Classifier for the Recognition of Persian/Arabic Handwritten Numerals , 2009, 2009 Seventh International Conference on Advances in Pattern Recognition.