Feature Extraction Techniques for Marathi Character Classification using Neural Networks Models

Main Article Content

Ms. Smita Ashokrao Bhopi, Mr. Manu Pratap Singh

Abstract

Hand written Marathi Character Recognition is challenges to the researchers due to the complex structure. This paper presents a novel approach for recognition of unconstrained handwritten Marathi characters. The recognition is carried out using multiple feature extraction methods and classification scheme. The initial stages of feature extraction are based upon the pixel value features and the classification of the characters is done according to the structural parameters into 44 classes. The final stage of feature extraction makes use of the zoning features. First Pixel values are used as features and these values are further modified as another set of features. All these features are then applied to neural network for recognition. A separate neural network is built for each type of feature. The average recognition rate is found to be 67.96% , 82.67%,63,46% and 76.46% respectively for feed forward , radial basis , elman and pattern recognition neural networks for handwritten marathi characters.

Article Details

How to Cite
, M. S. A. B. M. M. P. S. (2018). Feature Extraction Techniques for Marathi Character Classification using Neural Networks Models. International Journal on Future Revolution in Computer Science &Amp; Communication Engineering, 4(6), 70–76. Retrieved from http://www.ijfrcsce.org/index.php/ijfrcsce/article/view/1690
Section
Articles