Journal Paper

Paper Title - Recognition Of Handwritten Devanagari Characters Using Machine Learning Approach


Abstract
Recognition of devanagri characters poses challenges to the researchers due to their complex structure. This paper presents a methodology for recognition of unconstrained handwritten Marathi characters. The recognition is carried out using array based feature extraction and artificial neural network classifier. Fifty handwritten characters from 10 people resulting 500 characters are used for experimentation. The handwritten characters are scanned, preprocessed and on every individual character feature extraction is applied. The feature vectors are given as an input for training to back propagation neural network. Testing is carried out using individual characters as well as sentences. The accuracy obtained for recognition of 50 individual handwritten characters is 92%. For handwritten sentences, recognition accuracy obtained is 88.25%. Keywords—Devanagri, OCR, ANN Classifier, Feature Extraction, Neural Network Model.


Author - Pankaj Kale, Arti V. Bang, Devashree Joshi

Citation - Pankaj Kale   ,   Arti V. Bang   ,   Devashree Joshi   ,   Pankaj Kale, Arti V. Bang, Devashree Joshi " Recognition Of Handwritten Devanagari Characters Using Machine Learning Approach " , International Journal of Industrial Electronics and Electrical Engineering , Volume-3, Issue-9  ( Sep, 2015 )

Indexed - Google Scholar


| PDF |
Viewed - 0
| Published on 2015-09-04