American Sign Language Recognition Using Multi-Dimensional Hidden Markov Models

Abstract

An American Sign Language (ASL) recognition system developed based on multidimensional Hidden Markov Models (HMM) is presented in this paper. A Cyberglove(superscript TM) sensory glove and a Flock of Birds(superscript ®) motion tracker are used to extract the features of ASL gestures. The data obtained from the strain gages in the glove defines the hand shape while the data from the motion tracker describes the trajectory of hand movement. Our objective is to continuously recognize ASL gestures using these input devices in real time. With the features extracted from the sensory data, we specify multi-dimensional states for ASL signs in the HMM processor. The system gives an average of 95% correct recognition for the 26 alphabets and 36 basic handshapes in the ASL after it has been trained with 8 samples. New gestures can be accommodated in the system with an interactive, learning processor. The developed system forms a sound foundation for continuous recognition of ASL full signs.