Sinhala Character Identification Using Orientation and Support Vector Machine

In this paper we proposed a method for printed Sinhala character recognition based on support vector classification and histogram oriented gradient descriptor (HOG descriptor). Converting the text on images into text is a challenging task. It becomes the most difficult task when considering local languages. Having
$400+$
characters in the Sinhala alphabet and the curve pattern of the letters causes some difficulties when converting the Sinhala text images into text. A histogram oriented gradient descriptor is used to extract the unique features of each individual character. A linear support vector machine is used to classify each character. Sinhala corpus was created to check the word correctness and permutation method used for the process of creating words.