A Deep Convolution Neural Network-Based SE-ResNext Model for Bangla Handwritten Basic to Compound Character Recognition
With the recent advancement in artificial intelligence, the demand for handwritten character recognition increases day by day due to its widespread applications in diverse real-life situations. As Bangla is the world’s 7th most spoken language, hence the Bangla handwritten character recognition is demanding. In Bangla, there are basic characters, numerals, and compound characters. Character identicalness, curviness, size and writing pattern variations, lots of angles, and diversity make the Bangla handwritten character recognition task challenging. Recently, few papers have been published which study Bangla numeral, basic, and handwritten compound characters and the accuracy level in all three areas. The main objective of this paper is to propose a novel model that performs equally outstanding in all three different character types and to increase the efficiency of building a real-world Bangla handwritten character recognition system. This work describes a novel method for recognizing Bangla basic to compound character using a very special deep convolutional neural network model known as Squeeze-and-Excitation ResNext. The architectural novelty of our model is in introducing the Squeeze and Excitation (SE) Block, a very simple mathematical block with simple computation but very effective in finding complex features. We obtained 99.80% accuracy from a benchmark dataset of Bangla handwritten basic, numerals, and compound characters containing 160,000 samples. Additionally, our model demonstrates outperforming results compared to other state-of-the-art models.
Keywords: Bangla handwritten-character recognition, Deep Convolutional Neural Network, Squeeze and Excitation ResNext, global average pooling.
KIBRIA M R, AHMED A, FIRDAWSI Z, et al. Bangla Compound Character Recognition using Support Vector Machine (SVM) on Advanced Feature Sets [C]. IEEE Region 10 Symposium (TENSYMP), Dhaka, Bangladesh, 2020.
PRAMANIK R, & BAG S. Shape decomposition-based handwritten compound character recognition for Bangla OCR [J]. Journal of Visual Communication and Image Representation, 2018 (50):123-134.
RABBY A S A, HAQUE S, ISLAM M S, et al. BornoNet: Bangla Handwritten Characters Recognition Using Convolutional Neural Network [C]. Proceedings of the International Conference on Advances in Computing and Communication (ICACC-2018), Kochi, India, 2018(9).
ALOM M Z, SIDIKE P, HASAN M, et al. Handwritten Bangla Character Recognition Using the State-of-the-Art Deep Convolutional Neural Networks [J]. Computational Intelligence and Neuroscience, Hindawi. Article ID 6747098, 2018, (2018): 13.
NASIB A U, KABIR H, AHMED R, et al. A Real Time Speech to Text Conversion Technique for Bengali Language [C]. 2018 Proceedings of the International Conference on Computer, Communication, Chemical, Material and Electronic Engineering (IC4ME2), Rajshahi, Bangladesh, 2018.
CHOWDHURY A R, BISWAS A, HASAN S M F, et al. Bengali Sign language to text conversion using artificial neural network and support vector machine [C]. 2017 3rd Proceedings of the International Conference on Electrical Information and Communication Technology (EICT), Khulna, Bangladesh, 2017.
HASAN M J, WAHID M F, & ALOM M S. Bangla Compound Character Recognition by Combining Deep Convolutional Neural Network with Bidirectional Long Short-Term Memory [C]. Proceedings of the International Conference on Electrical Information and Communication Technology (EICT), Khulna, Bangladesh, 2019.
REZA S, AMIN O B, & HASHEM M M A. Basic to Compound: A Novel Transfer Learning Approach for Bengali Handwritten Character Recognition [C]. Proceedings of the International Conference on Bangla Speech and Language Processing (ICBSLP), Sylhet, Bangladesh, 2019.
KESERWANI P, ALI T, & ROY P P. A two phase trained Convolutional Neural Network for Handwritten Bangla Compound Character Recognition [C]. Proceedings of the 9th International Conference on Advances in Pattern Recognition (ICAPR), Bangalore, India, 2017.
ALIF M A R, AHMED S, & HASAN M A. Isolated Bangla Handwritten Character Recognition with Convolutional Neural Network [C]. Proceedings of the International Conference of Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 2017.
KHAN M M, UDDIN M S, PARVEZ M Z, et al. A squeeze and excitation ResNeXt-based deep learning model for Bangla handwritten compound character recognition [J]. Journal of King Saud University – Computer and Information Sciences, 2021: 1-9.
LI ZH, WU Q, XIAO Y, et al. Deep Matching Network for Handwritten Chinese Character Recognition [J]. Pattern Recognition, 2020 (107): 107471, [EB/OL]. https://doi.org/10.1016/j.patcog.2020.107471.
GAN J, WANG W, & LU K. Compressing the CNN architecture for in-air handwritten Chinese character recognition [J]. Pattern Recognition Letters, 2020 (129): 190-197, [EB/OL]. https://doi.org/10.1016/j.patrec.2019.11.028
ELTAY M, ZIDOURI A, & AHMAD I. Exploring Deep Learning Approaches to Recognize Handwritten Arabic Texts [J]. IEEE Access, 2020 (8): 89882 - 89898, [EB/OL]. https://ieeexplore.ieee.org/document/9091836
BHAGYASREE P V, JAMES A, & SARAVANAN C. A Proposed Framework for Recognition of Handwritten Cursive English Characters using DAG-CNN [C]. Proceedings of the International Conference on Innovations in Information and Communication Technology (ICIICT), CHENNAI, India, 2019: 1-6.
SAUFI M M, ZAMANHURI M A, MOHAMMAD N, et al. Deep Learning for Roman Handwritten Character Recognition [J]. Indonesian Journal of Electrical Engineering and Computer Science, 2018, 12(2): 455-460.
CLANUWAT T, LAMB A, & KITAMOTO A. KuroNet: Pre-Modern Japanese Kuzushiji Character Recognition with Deep Learning [C]. Proceedings of the International Conference on Document Analysis and Recognition (ICDAR2019), Sydney, Australia, 2019: 607-614.
JANGID M, & SRIVASTAVA S. Handwritten Devanagari Character Recognition Using Layer-Wise Training of Deep Convolutional Neural Networks and Adaptive Gradient Methods [J]. Journal of Imaging, MDPI, 2018, 4(2): 41.
HUSNAIN M, MISSEN M M S, Mumtaz S, et al. Recognition of Urdu Handwritten Characters Using Convolutional Neural Network [J]. Applied Sciences, MDPI, 2019, 9, (13): 2758.
AMIN M S, YASIR S M, & AHN H. Recognition of Pashto Handwritten Characters Based on Deep Learning [J]. Sensors, MDPI, 2020, 20 (20): 5884.
PAREEKM J, SINGHANIA D, KUMARI R R, et al. Gujarati Handwritten Character Recognition from Text Images [J]. Procedia Computer Science, 2019, 171: 514-523.
FARDOUS A, AFROGE S. Handwritten Isolated Bangla Compound Character Recognition [C]. Proceedings of the International Conference on Electrical, Computer and Communication Engineering (ECCE), Cox’s Bazar, Bangladesh, 2019.
ASHIQUZZAMAN A K M, TUSHAR A K, DUTTA S, et al. An Efficient Method for Improving Classification Accuracy of Handwritten Bangla Compound Characters using DCNN with Dropout and ELU [C]. Proceedings of the International Conference on Research in Computational Intelligence and Communications Networks (ICRCICN), Dhaka, Bangladesh, 2017.
CHATTERJEE S, DUTTA R K, GANGULY D, et al. Bengali Handwritten Character Classification using Transfer Learning on Deep Convolutional Neural Network [C]. Proceedings of the International Conference on Intelligent Human Computer Interaction, Allahabad, India, 2019.
SAHA S, & SAHA N. A Lightning fast approach to classify Bangla Handwritten Characters and Numerals using newly structured Deep Neural Network [C]. Proceedings of the International Conference on Computational Intelligence and Data Science (ICCIDS 2018), Gurugram, India, 2018.
HU J, SHEN L, ALBANIE S, et al. Squeeze-and-Excitation Networks [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011-2023.
CLEVERT D A, UNTERTHINER T, HOCHREITER S. Fast and accurate deep network learning by exponential linear units (ELUs) [c]. 2016 Proceedings of the International Conference on Learning Representations. May 2, 2016 - May 4, 2016, San Juan, Puerto Rico [EB/OL]. arXiv:1511.07289, 2016.
MOHAMMED N, MOMEN S, ABEDIN A, et al. BanglaLekha-Isolated [EB/OL]. https://data.mendeley.com/datasets/hf6sf8zrkc/2, 2017.
LECUN Y, CORTES C, & BURGES C J. MNIST handwritten digit database. AT&T Labs. [EB/OL]. http://yann.lecun.com/exdb/mnist, 2010, 2.
THAKKAR V, TEWARY S, & CHAKRABORTY C. Batch Normalization in Convolutional Neural Networks-A comparative study with CIFAR-10 data [C]. Proceedings of the International Conference on Emerging Applications of Information Technology (EAIT), Kolkata, India, 2018.
- There are currently no refbacks.