OFFLINE INDIAN CHARACTER RECOGNITION USING SUPPORT VECTOR MACHINE (SVM) CLASSIFIER

This paper present Offline Indian Character Recognition using SVM classifier. In the sense of Indian character various script are comprises like Devanagari, Gurmukhi etc. This paper carried work on Devanagari script which include several language like Hindi, Bengali, etc which is also discussed in next section of this paper. In this research report an experimental summary on offline handwritten Devanagari character recognition using different feature extraction techniques like image centroid zone (ICZ), zone centroid zone (ZCZ) and combination of both methods namely Image centroid zone as well as zone centroid zone (ICZ and ZCZ) .After execution of the feature extraction techniques, image produced several feature vector, then different feature vectors are passed into support vector machine SVM Classifier for recognition. Different methods produced several recognition accuracy from ICZ , ZCZ and hybrid method are 77.51%, 98.54% and 77.71% accordingly.