Real Time Indian Sign Language Recognition Using SIFT by Depth Sensor

Abstract

Indian sign language (ISL) provides interface to hard hearing community to communicate with the society. The hand plays a significant role in any sign language. ISL is composed of static and dynamic sign hand gestures. In proposed system, depth sensor is used to track and segment hand in clutter background and feature extracted by scale-invariant feature transform (SIFT). Finally, we recognize 10 hand gestures which represent as number in ISL using multi support vector machine with an accuracy of 94.9%.

Keywords: Sign Language Recognition (ISL), Depth Sensor, Scale-Invariant Feature Transform (SIFT), Multi-Support Vector Machine (MSVM)