Hybrid feature extraction and optimized classification for handwritten Devanagari character recognition using ATOA and BO-SVM
Shubham Srivastava1, Ajay Verma2 and Shekhar Sharma3
Research Scholar, Institute of Engineering and Technology,Devi Ahilya Vishwavidyalaya, Indore,India1
Professor and Head , Department of Computer Science,Institute of Engineering and Technology, Devi Ahilya Vishwavidyalaya, Indore,India2
Professor, Electronics and Telecommunication Engineering,Shri Govindram Seksaria Institute of Technology and Science, Indore,India3
Professor and Head , Department of Computer Science,Institute of Engineering and Technology, Devi Ahilya Vishwavidyalaya, Indore,India2
Professor, Electronics and Telecommunication Engineering,Shri Govindram Seksaria Institute of Technology and Science, Indore,India3
Corresponding Author : Shubham Srivastava
Recieved : 14-March-2025; Revised : 04-May-2025; Accepted : 12-May-2025
Abstract
Handwritten character recognition (HCR) is a critical yet challenging task in the fields of computer vision and pattern recognition. It involves the automatic interpretation and classification of characters written by hand, which often vary significantly in style, orientation, and quality. This study proposes a comprehensive methodology aimed at significantly enhancing the accuracy of HCR. The approach comprises several integrated stages. Initially, precise row and column segmentation techniques are employed to isolate individual characters from the input images. A novel hybrid feature extraction method is then introduced, incorporating multiple techniques: speeded-up robust features (SURF) for local feature detection, Gabor filters for capturing texture and edge information, zone transformation for extracting region-specific features, and the radon transform for encoding orientation-related details. To further improve recognition performance, the arithmetic-trigonometric optimization algorithm (ATOA) is utilized for feature selection, enabling the identification of the most relevant features for classification. Subsequently, a support vector machine (SVM) classifier, optimized using Bayesian optimization (BO), is trained to achieve superior classification accuracy through fine-tuned hyperparameters. The proposed methodology demonstrates exceptional performance, with the BO-SVM classifier attaining a classification accuracy of 98.53%. These results highlight the effectiveness of the approach across various character recognition tasks, surpassing the performance of traditional techniques. In addition to its high accuracy, the proposed methodology offers a systematic framework to address the diverse challenges inherent in handwritten character recognition.
Keywords
Handwritten character recognition, Feature extraction, Support vector machine, Bayesian optimization, Arithmetic-trigonometric optimization.
Cite this article
Srivastava S, Verma A, Sharma S. Hybrid feature extraction and optimized classification for handwritten Devanagari character recognition using ATOA and BO-SVM. International Journal of Advanced Technology and Engineering Exploration. 2025;12(126):787-802. DOI : 10.19101/IJATEE.2024.111100295
References
[1]
Jaiswal S, Chaudhari K, Tade S, Khirwadkar S, Pande A. Empirical review on handwritten Devanagari script recognition techniques using AI approaches. International Journal of Creative Computing. 2024; 2(2):119-37.
[2]
Inunganbi S. A systematic review on handwritten document analysis and recognition. Multimedia Tools and Applications. 2024; 83(2):5387-413.
[3]
Sharma C, Sharma S, Sakshi, Chen HY. Advancements in handwritten Devanagari character recognition: a study on transfer learning and VGG16 algorithm. Discover Applied Sciences. 2024; 6(12):1-20.
[4]
Devendiran S, Nedungadi P, Raman R. Handwritten text recognition using VGG19 and HOG feature descriptors. In international conference on interdisciplinary approaches in technology and management for social innovation 2024 (pp. 1-4). IEEE.
[5]
El MM. An effective combination of convolutional neural network and support vector machine classifier for Arabic handwritten recognition. Automatic Control and Computer Sciences. 2023; 57(3):267-75.
[6]
Singh S, Garg NK, Kumar M. VGG16: offline handwritten Devanagari word recognition using transfer learning. Multimedia Tools and Applications. 2024; 83(29):72561-94.
[7]
Duarte JM, Berton L. A review of semi-supervised learning for text classification. Artificial Intelligence Review. 2023; 56(9):9401-69.
[8]
Singh S, Sharma A, Chauhan VK. Indic script family and its offline handwriting recognition for characters/digits and words: a comprehensive survey. Artificial Intelligence Review. 2023; 56(Suppl 3):3003-55.
[9]
Jabde KM, Patil HC, Vibhute DA, Mali S. A comprehensive literature review on air-written online handwritten recognition. International Journal of Computing and Digital Systems. 2024; 15(1):307-22.
[10]
Sharma A, Mithun BN. Deep learning character recognition of handwritten Devanagari script: a complete survey. In international conference on contemporary computing and communications 2023 (pp. 1-6). IEEE.
[11]
Nagane AS, Patil CH, Mali SM. Classification of brahmi script characters using HOG features and multiclass error-correcting output codes (ECOC) model containing SVM binary learners. In international conference on intelligent and innovative technologies in computing, electrical and electronics 2023 (pp. 448-51). IEEE.
[12]
Kaur R. An overview of advanced technologies applied to identified printed and handwritten text in Gurmukhi script: a review. Journal Punjab Academy of Sciences. 2023; 23:194-207.
[13]
Bose SR, Singh R, Joshi Y, Marar A, Regin R, Rajest SS. Light weight structure texture feature analysis for character recognition using progressive stochastic learning algorithm. In advanced applications of generative AI and natural language processing models 2024 (pp. 144-58). IGI Global Scientific Publishing.
[14]
Rajpal D, Garg AR. Deep learning model for recognition of handwritten Devanagari numerals with low computational complexity and space requirements. IEEE Access. 2023; 11:49530-9.
[15]
Moudgil A, Singh S, Sareen B, Wadhwa S. Devanagari ancient manuscript recognition using AlexNet. In 11th international conference on reliability, infocom technologies and optimization (Trends and Future Directions) 2024 (pp. 1-5). IEEE.
[16]
Narang S, Jindal MK, Kumar M. Devanagari ancient documents recognition using statistical feature extraction techniques. Sādhanā. 2019; 44(6):141.
[17]
Khanderao MS, Ruikar S. Character segmentation and recognition of Indian Devanagari script. In ICT analysis and applications: proceedings of ICT4SD 2019 (pp. 529-37). Springer Singapore.
[18]
Dessai B, Patil A. A deep learning approach for optical character recognition of handwritten Devanagari script. In 2nd international conference on intelligent computing, instrumentation and control technologies 2019 (pp. 1160-5). IEEE.
[19]
Singh S, Aggarwal A, Dhir R. Use of gabor filters for recognition of handwritten Gurmukhi character. International Journal of Advanced Research in Computer Science and Software Engineering. 2012; 2(5):234-40.
[20]
Rani R, Dhir R, Lehal GS. Gabor features based script identification of lines within a bilingual/trilingual document. International Journal of Advanced Science and Technology. 2014; 66:1-12.
[21]
Diem M, Sablatnig R. Recognizing characters of ancient manuscripts. In computer vision and image analysis of Art 2010 (pp. 35-46). SPIE.
[22]
Helmer S, Lowe DG. Object class recognition with many local features. In conference on computer vision and pattern recognition workshop 2004 (pp. 187-7). IEEE.
[23]
Reddy RV, Babu UR. Handwritten Hindi character recognition using deep learning techniques. International Journal of Computer Sciences and Engineering. 2019; 7(2):1-17.
[24]
Gadekallu TR, Khare N, Bhattacharya S, Singh S, Maddikunta PK, Ra IH, et al. Early detection of diabetic retinopathy using PCA-firefly based deep learning model. Electronics. 2020; 9(2):1-16.
[25]
Puri S, Singh SP. An efficient Devanagari character classification in printed and handwritten documents using SVM. Procedia Computer Science. 2019; 152:111-21.
[26]
Bhunia AK, Konwer A, Bhunia AK, Bhowmick A, Roy PP, Pal U. Script identification in natural scene image and video frames using an attention based convolutional-LSTM network. Pattern Recognition. 2019; 85:172-84.
[27]
Tian S, Bhattacharya U, Lu S, Su B, Wang Q, Wei X, et al. Multilingual scene character recognition with co-occurrence of histogram of oriented gradients. Pattern Recognition. 2016; 51:125-34.
[28]
Shi B, Yao C, Zhang C, Guo X, Huang F, Bai X. Automatic script identification in the wild. In 13th international conference on document analysis and recognition 2015 (pp. 531-5). IEEE.
[29]
Aneja N, Aneja S. Transfer learning using CNN for handwritten Devanagari character recognition. In 1st international conference on advances in information technology 2019 (pp. 293-6). IEEE.
[30]
Huang G, Liu Z, Van DML, Weinberger KQ. Densely connected convolutional networks. In proceedings of the conference on computer vision and pattern recognition 2017 (pp. 4700-8). IEEE.
[31]
Talebi H, Milanfar P. Learning to resize images for computer vision tasks. In proceedings of the international conference on computer vision 2021 (pp. 497-506). IEEE/CVF.
[32]
Singh N. An efficient approach for handwritten Devanagari character recognition based on artificial neural network. In 5th international conference on signal processing and integrated networks 2018 (pp. 894-7). IEEE.
[33]
Jadhav S, Inamdar V. Convolutional neural network and histogram of oriented gradient based invariant handwritten Modi character recognition. Pattern Recognition and Image Analysis. 2022; 32(2):402-18.
[34]
Arora S, Bhatcharjee D, Nasipuri M, Malik L. A two stage classification approach for handwritten Devnagari characters. In international conference on computational intelligence and multimedia applications 2007 (pp. 399-403). IEEE.
[35]
Ghimire A, Chapagain A, Bhattarai U, Jaiswal A. Nepali handwriting recognition using convolution neural network. International Research Journal of Innovations in Engineering and Technology. 2020; 4(5):1-5.
[36]
Gharde SS, Ramteke RJ, Kotkar VA, Bage DD. Handwritten Devanagari numeral and vowel recognition using invariant moments. In international conference on global trends in signal processing, information computing and communication 2016 (pp. 255-60). IEEE.
[37]
Kundaikar T, Fadte S, Karmali R, Pawar JD. Automatic Hindi OCR error correction using MLM-BERT. Ingénierie des Systèmes d’Information. 2024; 29(2): 619-26.
[38]
Omayio EO, Indu S, Panda J. Word spotting and character recognition of handwritten Hindi scripts by integral histogram of oriented displacement (IHOD) descriptor. Multimedia Tools and Applications. 2024; 83(1):1379-406.
[39]
Bisht M, Gupta R. Offline handwritten Devanagari modified character recognition using convolutional neural network. Sādhanā. 2021; 46:1-4.
[40]
Wan CH, Lee LH, Rajkumar R, Isa D. A hybrid text classification approach with low dependency on parameter by integrating K-nearest neighbor and support vector machine. Expert Systems with Applications. 2012; 39(15):11880-8.
[41]
Mehta N, Doshi J. Segmentation methods: a review. International Journal for Research in Applied Science and Engineering Technology. 2020; 8:536-40.
[42]
Panneer SAM, Hussin FA, Ibrahim R, Bingi K. Arithmetic-trigonometric optimization algorithm. In optimal fractional-order predictive PI controllers: for process control applications with additional filtering 2022 (pp. 99-133). Singapore: Springer Nature Singapore.
[43]
Deore SP, Pravin A. Devanagari handwritten character recognition using fine-tuned deep convolutional neural network on trivial dataset. Sādhanā. 2020; 45(1):243.
[44]
Gaikwad SS, Nalbalwar SL, Nandgaonkar AB. Devanagari handwritten characters recognition using DCT, geometric and hue moments feature extraction techniques. Sādhanā. 2022; 47(3):112.
[45]
Sahare P, Dhok SB. Multilingual character segmentation and recognition schemes for Indian document images. IEEE Access. 2018; 6:10603-17.