Analysis of Proposed Technique for Graphical Representation of Isolated Character Using SVM Classifier

A graphical representation is a visual display of data and statistical results. It is more often and effective than presenting data in tabular form. Optical Character Recognition (OCR) requires a graphical representation of text to interpret, which usually comes from a scanned image. Support Vector Machine (SVM) describes the concept that how the decision planes are made which helps in defining the decision boundaries. In this paper a method of isolated graphical representation has been proposed using SVM Classifier. The performance is measured in the terms of accuracy using different font styles and font sizes. The work is done on Sindhi Character Set. The result shows the accuracy recognition rate achieved with SVM Classifier is much better than existing Global Transformation and Feature Extraction Techniques.