A Hybrid Gesture Recognition Method for American Sign Language

Abstract

Gesture based communication is a method of correspondence between the ordinary and hard of hearing people in which the vision based procedure is utilized. This paper proposes a novel methodology of hand gesture recognition system for American Sign Language (ASL), which will perceive communication via gestures signals in an ongoing situation. A hybrid based descriptor, which joins the benefits of LBP (Local binary pattern), SP (super pixels) and SURF (Speeded Up Robust Features) strategies, is utilized as a consolidated list of capabilities to accomplish a improved identification rate beside among a little moment in time computational difficulty. In additional increase the detection speed and create the appreciation framework strong to view-point varieties, the idea of derived features from the accessible list of capabilities is presented. K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) are utilized for hybrid arrangement of single marked letter. Comparative investigation of these strategies with other well known methods demonstrates that the constant proficiency and robustness are better. The performances parameters will be used in this method are accuracy, sensitivity, precision, FNR and FDR. Keywords: American Sign Language, FNR, FDR, Hand Gesture Recognition, KNN, LBP, Recognition Rate, SP, SURF, SVM