Recognition of Spatiotemporal Gestures in Sign Language Using Gesture Threshold HMMs

Abstract

In this paper, we propose a framework for the automatic recognitionof spatiotemporal gestures in Sign Language. We implement an extensionto the standard HMM model to develop a gesture threshold HMM (GT-HMM)framework which is specifically designed to identify inter gesturetransitions. We evaluate the performance of this system, and differentCRF systems, when recognizing gestures and identifying inter gesturetransitions. The evaluation of the system included testing the performanceof conditional random fields (CRF), hidden CRF (HCRF) and latent-dynamicCRF (LDCRF) based systems and comparing these to our GT-HMM basedsystem when recognizing motion gestures and identifying inter gesturetransitions.