Accuracy Improvement of Off-line Handwritten Tamil Character Recognition
Authors :
Punitharaja .Ka,* and Dr. P. Elangob
Author Address :
aDepartment of Computer Science, T.B.M.L. College, PORAYAR 609307, Nagapattinam District, Tamilnadu, India
bPerunthalaivar Kamarajar Institute of Engineering and Technology (PKIET), Nedungadu, KARAIKAL-609603, Puducherry, India
*Corresponding author.
Abstract :
India is a multi-lingual and multi-script country, where eighteen official scripts are accepted and have over hundred regional languages. Tamil is the most popular language in the world and particularly in Tamilnadu. More than 8 crore Tamils live in Tamil Nadu and Pondicherry. About one crore Tamils live in the other states of India. Outside India, Sri Lanka, Burma, Malaysia, Singapore, Indonesia, South Africa, Fiji, Mauritius islands are some of the countries having a large number of Tamil speaking people. Thus, the work on Tamil character is very useful for the Tamil community around the world.
This paper deals with the recognition of off-line handwritten Tamil characters. Tamil Character recognition is a most challenging task in image processing and pattern recognition fields. Handwritten character recognition has received extensive attention in academic and production fields. Off-line handwriting recognition is the subfield of optical character recognition. Here two sets of feature are computed and two classifiers are combined to get higher accuracy of Tamil character recognition. First feature set is computed based on the directional information obtained from the arc tangent of the gradient. Since most of the Tamil handwritten characters have some curve-like parts, curvature-based feature guided by gradient information is computed for the second set of features. Curvature feature detection calculated using bi-quadratic interpolation method. Combined use of Support Vector Machines (SVM) and Modified Quadratic Discriminant Function (MQDF) are applied here for better performance of Tamil character recognition.
Keywords :
Handwritten character recognition, Tamil characters, Gradient feature, Curvature feature, Support Vector Machines (SVM), Quadratic Discriminant Function (MQDF)
DOI :
Article Info :
Received : August 12, 2015; Accepted : September 30, 2015.
