The main aim of this work is the classification of Tamil palm leaf manuscript segmented characters using Machine Learning approach. For the segmentation of characters, first the images of the palm leaf manuscripts were allowed to preprocessing which includes filtering and thresholding. After the preprocessing stage, the preprocessed images were allowed for character segmentation using contour based bounding box segmentation. Then the segmented Tamil palm leaf manuscript characters were labelled with different classes for classification. To classify the characters Adaptive Backpropagation Neural Network (ABPN) with Shannon activation function was used with Nguyen Widrow weight initialization. For neural network, normally we use random initialization to generate the weights. Rather than random initialization here Nguyen-Widrow weight initialization technique was implemented. For comparison ABPN with Shannon activation function (method 1) and ABPN with Shannon activation function using Nguyen-Widrow initialization was used, from this ABPN with Shannon activation function using Nguyen-Widrow gives 96% of accuracy for Tamil palm leaf character classification.

ABPN, Bounding box, Convex hull, Contour, Shannon, Machine Learning

[1] Amit Choudhary, Rahul Rishi, Savita Ahlawat, “Off-Line Handwritten Character Recognition using Features Extracted from Binarization Techniques”, AARSI Procedia, , Vol.4, pp.306-312, 2013.
[2] Dino Neinhold, Kilian Schwab, Rolf Dornberger and Thomas Hanne, “Effects of Weight Initialization in a Feedforward Neural Network for Classification Using a Modified Genetic Algorithm”, IEEE, pp.6-12, 2015.
[3] Hossin. M., Sulaiman. M.N, “A Review on Evaluation Metrics for Data Classification Evaluations”, International Journal Of Data Mining & Knowledge Management Process, Vol.5, No.2, pp.1-11, 2015.
[4] JeyaseeliSubavathi S, Kathirvalavakumar T, “Adaptive modified backpropagation algorithm based on differential errors”, International Journal of Computer Science, Engineering and Applications (IJCSEA),; Vol.1, pp. 21-34, 2011.
[5] Karthigaiselvi. M, Kathirvalavakumar. T, “Structural Run Based Feature Vector to Classify Printed Tamil Characters using Neural Network”, International Journal of Engineering Research and Application, Vol.7, Issue.7, pp.44-63, 2017.
[6] KavithaSubramani, Murugavalli. S, “A Novel Binarization Method for Degraded Tamil Palm Leaf Images”, IEEE, pp.176-181, 2016.
[7] Kiruba. B, Nivethitha. A, Vimaladevi. M, “Segmentation of Handwritten Tamil Character from Palm Script using Histogram Approach”, International Journal of Informative and Futuristic Research, Vol.4,Issue.5, pp.6418-6424,2017.
[8] NarahariSastryPanyam, Vijaya Lakshmi. T.R, Ramakrishnan Krishnan, Koteswara Rao. N.V, “Modeling of Palm Leaf Character Recognition System using Transform based Techniques”, Pattern Recognition Letters, Vol.84, pp.29-34,2016.
[9] Nibaran Das, Jagan Mohan Reddy, Ram Sarkar, SubhadipBasu, MahantapasKundu, MitaNasipuri, Dipak Kumar Basu, “A Statistical-Topological Feature Combination for Recognition of Handwritten Numerals”, Applied Soft Computing ,Vol.12, pp.2486-2495 , 2012.
[10] Sornam.M, Muthu SubashKavitha, Poornima Devi.M, “An Efficient Morlet Function based Adaptive Method for Faster Backpropagation for Handwritten Character Recognition”, IEEE, pp.135-139, 2016.
[11] Sornam.M, Poornima Devi. M, “Tamil Palm Leaf Manuscript Character Segmentation using GLCM Feature Extraction”, International Journal of Computer Science and Engineering, Vol.6, Issue.4,pp.167-173,2018.
[12] Sornam. M, C. Vishnu Priya, “Deep Convolutional Neural Network for Handwritten Tamil Character Recognition Using Principal Component Analysis”, Smart and Innovative Trends in Next Generation Computing Technologies, pp. 778-787, 2018.
[13] Tien-Chien Chang and Shu-Yuan Chen, “Character Segmentation using Convex hull Techniques”, International Journal of Pattern Recognition and Artificial Intelligence, Vol.13, No.6, pp.833-858,1999.
[14] Vellingiriraj. E.K, Balasubramanie. P, “Recognition of Ancient Tamil Handwritten Characters in Historical Documents by Boolean Matrix and BFS Graph”, International Journal of Computer Science and Technology, Vol.5, pp.65-68,2014.
[15] Vijaya Lakshmi. T.R, “Reduction of Features to Identify Characters from Degraded Historical Manuscripts”, Alexandria Engineering Journal, pp.1-7, 2017.
[16] YapingZang, Shan Liang, ShuaiNie, Wenju Liu, Shouye Peng, “Robust Offline Handwritten Character Recognition through Exploring Writer-Independent features under the guidance of Printed data”, Pattern Recognition Letters, Vol.106, pp.20-26,2018.
[17] Yi-Chou Wu, Fei-Yin, Cheng-Lin Liu, “Improving Handwritten Chinese Text Recognition using Neural Network Language Models and Convolutional Neural Network Shape Models”, Pattern Recognition, Vol.65,pp.251-264, 2017.