A Static Tamil Sign Language Recognition System

Abstract: Sign language recognition is a very challenging research area. In this paper, a system to recognize static gestures representing Tamil words has been proposed. The recognition of human gestures and facial expressions in image sequences is an important and challenging problem that enables a host of human-computer interaction applications. In this work a new view for recognizing sign language has been proposed. Most researches on continuous sign language recognition were done with frames obtained by processing the videos with regular/equal interval. If a system developed is strong enough for processing the static gestures then it would be the finest system to process the frames obtained while processing the continuous gestures. This work contains three phases of work. First phase is preprocessing, in which the obtained images are processed through the steps like resize, gray conversion, filtering for reducing the distortion and Black and White conversion. Black and white image is taken purposely so that shape descriptors can be applied to extract the required features. Region-based analysis exploits both boundary and interior pixels of an object. Solidity, Perimeter, Convex area, Major axis length, Minor axis length, Eccentricity, Orientation are some of the shape descriptors used as features in this work. Proximal Support Vector Machine classifier has been considered for classification which provides a good result with less computation time for larger datasets. The features derived are used to train the binary classifier first, secondly the testing images has been introduced for classification. Since, we handled a binary classifier we performed a one-versus-all kind of classification. A The proposed Sign Language Recognition System is able to recognize images with 91% accuracy.

Keywords: Sign Language recognition, Shape Descriptors, Machine learning, Proximal Support Vector Machine