National conference on Digital Image and Signal Processing |
Foundation of Computer Science USA |
DISP2015 - Number 3 |
April 2015 |
Authors: C. H. Patil, S. M. Mali |
fff782a0-722d-425c-ba26-1ce3dfe56051 |
C. H. Patil, S. M. Mali . Recognition of Isolated Handwritten Marathi Vowels using Symmetric Density and Moment Invariant Features. National conference on Digital Image and Signal Processing. DISP2015, 3 (April 2015), 27-31.
In this paper, combination of zone based symmetric density feature and moment invariant feature is proposed for recognition of isolated handwritten Marathi vowels. Recognition of handwritten Marathi vowels is a challenging task due to their interclass structural similarities. Since a standard database does not exist for handwritten Marathi vowels, as a part of this work database of 2294 handwritten Marathi vowels was developed. Pre-processing techniques are applied to remove noise and there zone based symmetric density features are extracted. In addition to zone based symmetric density feature, moment invariants for each image is extracted. Proposed methodology is tested using fivefold cross validation technique; maximum 92. 98 percent recognition accuracy was noted for fold III using SVM classifier. The recognition accuracy was compared using k-NN and SVM classifiers. Average recognition rate achieved 88. 17 percent and 86. 93 percent using k-NN classifier and SVM classifier respectively.