Similar Shaped Ancient Tamil Character Recognition Using Gradient Feature and Template Matching

Recognition of similar shaped ancient Tamil character is a difficult problem and in character recognition system most of the errors occur from similar shaped characters. Researchers have used many methods of feature extraction for ancient Tamil character recognition. In this paper we proposed a novel feature extraction technique to improve the recognition results of two similar shaped ancient Tamil characters. In this work, we have taken the samples of offline ancient Tamil characters from six different centuries. The technique is based on F-ratio (Fisher Ratio), a statistical measure defined by the ratio between-class variance and within-class variance. F-ratio modifies the feature vector of two similar shape characters by weighting the feature elements. This weighting scheme enhances the feature elements that belongs to the distinguishable portions of the similar shaped characters and reduces the feature elements of the common portion of the characters, so that similar shaped characters can be identified easily. We considered pair of similar shape ancient characters of different stone inscriptions and we noted that f-ratio based feature weighting shows better recognition results. The proposed system achieves a maximum recognition accuracy of 94% using gradient features and template matching method.