Abstract
Handwritten word recognition, a classical pattern recognition problem, converts a word image into its machine editable form. Mainly two basic approaches are followed to solve this problem, one is segmentation-based and the other is holistic. A number of research attempts have shown that the holistic approach performs better than its counterpart when the lexicon is predefined, fixed and small in size. Relying on this, initial benchmark recognition accuracy on CMATERdb2.1.2, a publicly available database consists of handwritten city names in Bangla, was reported following a holistic word recognition protocol. In the present work, we have followed the same trend to recognize the word samples of the said database and set a new benchmark recognition accuracy. A sparse convolutional neural network (CNN)-based model which is a low-cost trainable model has been developed for this. We have relied on a recently proposed hypothesis, known as lottery ticket hypothesis for pruning the layers of CNN model methodically, and derived a low-resource model having much less number of training parameters. This model competently surpasses the previously reported recognition accuracy on the said database by a significant margin with an axed training cost.
Similar content being viewed by others
References
Liu C-L, Yin F, Wang D-H, Wang Q-F (2013) Online and offline handwritten Chinese character recognition: benchmarking on new databases. Pattern Recognit 46:155–162
Malakar S, Ghosh P, Sarkar R et al (2011) An improved offline handwritten character segmentation algorithm for Bangla script. In: Proceedings of the 5th Indian international conference on artificial intelligence, IICAI 2011
Kirn G, Govindaraju V (1997) A lexicon driven approach to handwritten word recognition for real-time applications. IEEE Trans Pattern Anal Mach Intell 19:366–379. https://doi.org/10.1109/34.588017
Edelman S, Flash T, Ullman S (1990) Reading cursive handwriting by alignment of letter prototypes. Int J Comput Vis 5:303–331. https://doi.org/10.1007/BF00126503
Morita M, Sabourn R, El Yacoubi A et al (2001) Handwritten month word recognition on Brazilian Bank Checks. In: Sixth international conference on document analysis and recognition. IEEE, pp 972–976
Namane A, Guessoum A, Meyrueis P (2005) New holistic handwritten word recognition and its application to French legal amount. In: International conference on pattern recognition and image analysis. Springer, pp 654–663
Al Aghbari Z, Brook S (2009) HAH manuscripts: a holistic paradigm for classifying and retrieving historical Arabic handwritten documents. Expert Syst Appl 36:10942–10951. https://doi.org/10.1016/j.eswa.2009.02.024
Malakar S, Ghosh M, Sarkar R, Nasipuri M (2018) Development of a two-stage segmentation-based word searching method for handwritten document images. J Intell Syst. https://doi.org/10.1515/jisys-2017-0384
Bhowmik S, Malakar S, Sarkar R et al (2019) Off-line Bangla handwritten word recognition: a holistic approach. Neural Comput Appl 31:5783–5798
Frankle J, Carbin M (2019) The lottery ticket hypothesis: finding sparse, trainable neural networks. In: International conference on learning representations
El Qacimy B, Kerroum MA, Hammouch A (2016) Word-based Arabic handwritten recognition using SVM classifier with a reject option. In: International conference on intelligent systems design and applications, ISDA. IEEE, pp 64–68
Dasgupta J, Bhattacharya K, Chanda B (2016) A holistic approach for Off-line handwritten cursive word recognition using directional feature based on Arnold transform. Pattern Recognit Lett 79:73–79. https://doi.org/10.1016/j.patrec.2016.05.017
Malakar S, Sharma P, Singh PK et al (2017) A holistic approach for handwritten Hindi Word recognition. Int J Comput Vis Image Process 7:59–78. https://doi.org/10.4018/IJCVIP.2017010104
Bhowmik S, Polley S, Roushan MG et al (2015) A holistic word recognition technique for handwritten Bangla words. Int J Appl Pattern Recognit 2:142–159
Barua S, Malakar S, Bhowmik S, et al (2017) Bangla handwritten city name recognition using gradient-based feature. In: 5th international conference on frontiers in intelligent computing: theory and applications. Springer, Singapore, pp 343–352
Tamen Z, Drias H, Boughaci D (2017) An efficient multiple classifier system for Arabic handwritten words recognition. Pattern Recognit Lett 93:123–132. https://doi.org/10.1016/j.patrec.2017.01.020
Sahoo S, Nandi SK, Barua S et al (2018) Handwritten Bangla word recognition using negative refraction based shape transformation. J Intell Fuzzy Syst. https://doi.org/10.3233/JIFS-169712
Ghosh M, Malakar S, Bhowmik S et al (2019) Feature selection for handwritten word recognition using memetic algorithm. In: Advances in intelligent computing. Springer, New York, pp 103–124
Dehghan M, Faez K, Ahmadi M, Shridhar M (2001) Handwritten Farsi (Arabic) word recognition: a holistic approach using discrete HMM. Pattern Recognit 34:1057–1065. https://doi.org/10.1016/S0031-3203(00)00051-0
Shaw B, Parui SK, Shridhar M (2008) Offline handwritten devanagari Word recognition: a holistic approach based on directional chain code feature and HMM. In: 2008 international conference on information technology. IEEE, pp 203–208
Bhowmik TK, Parui SK, Roy U (2008) Discriminative HMM training with GA for handwritten word recognition. In: 2008 19th international conference on pattern recognition. IEEE, pp 1–4
Malakar S, Ghosh M, Bhowmik S et al (2019) A GA based Hierarchical Feature Selection Approach for Handwritten Word Recognition. Neural Comput Appl. https://doi.org/10.1007/s00521-018-3937-8
Roy PP, Bhunia AK, Das A et al (2016) HMM-based Indic handwritten word recognition using zone segmentation. Pattern Recognit 60:1057–1075. https://doi.org/10.1016/j.patcog.2016.04.012
Pechwitz M, Maddouri SS, Märgner V (2002) IFN/ENIT-database of handwritten Arabic words. In: Proceedings of CIFED. Citeseer, pp 127–136
Centre for Pattern Recognition and Machine Intelligence. http://www.concordia.ca/research/cenparmi.html. Accessed 17 Dec 2018
IAM Handwriting Database. http://www.fki.inf.unibe.ch/databases/iam-handwriting-database. Accessed 17 Dec 2018
Bhowmik S, Malakar S, Sarkar R, Nasipuri M (2014) Handwritten Bangla word recognition using elliptical features. In: Proceedings—2014 6th international conference on computational intelligence and communication networks, CICN 2014
Majumdar A Lottery_Ticket_Hypothesis-TensorFlow_2. https://github.com/arjun-majumdar/Lottery_Ticket_Hypothesis-TensorFlow_2/blob/master/tfmot_sparsity_experiment.ipynb. Accessed 7 Feb 2020
Liu L, Chen J, Fieguth P et al (2019) From BoW to CNN: two decades of texture representation for texture classification. Int J Comput Vis 127:74–109
Lin T-Y, Maji S (2016) Visualizing and understanding deep texture representations. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp 2791–2799
Vasudev R Understanding and Calculating the number of Parameters in Convolution Neural Networks (CNNs). https://towardsdatascience.com/understanding-and-calculating-the-number-of-parameters-in-convolution-neural-networks-cnns-fc88790d530d. Accessed 15 Nov 2019
Kundu S, Paul S, Singh PK et al Understanding NFC-Net: a deep learning approach to word-level handwritten Indic script recognition. Neural Comput Appl. https://doi.org/10.1007/s00521-019-04235-4
Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems, pp 1097–1105
Acknowledgements
We would like to thank CMATER research laboratory of the Computer Science and Engineering Department, Jadavpur University, India, for providing us the infrastructural support. This work is partially supported by the PURSE-II and UPE-II, Jadavpur University projects. Ram Sarkar is thankful to DST, Govt. of India, for the Grant (EMR/2016/007213) to carry out this research.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
We declare that we have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Malakar, S., Paul, S., Kundu, S. et al. Handwritten word recognition using lottery ticket hypothesis based pruned CNN model: a new benchmark on CMATERdb2.1.2. Neural Comput & Applic 32, 15209–15220 (2020). https://doi.org/10.1007/s00521-020-04872-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-020-04872-0