Development of an adaptive hand gesture database: Motion trajectories cue

Abstract

The developments of computer technology become more fascinated especially through human-computer interaction (HCI). The visual interpretations and analysis of the human body motion are the intermediary of interaction between human and computer. In particular, hand gesture is used as a non-verbal tool for human to communicate with machines that pertains to the use of human gesture, especially hand as a medium of interaction between them. Recently, many studies for recognizing the human gesture have been reported, and most of them deal with the shape and movement of hands. Moreover, they also discuss on factor that contributed to the effect of individual factors in arm motions. This paper mainly concentrated on the development of a gesture database to eliminate individual factors which affect the efficiency of the recognition system. Several performers were selected with several physical characteristic for the classification purposes. Experimental data were collected using geometrical gesture perform by the performers. An adaptive gesture recognition system is proposed, and the system could adaptively select the correspond database for the purpose of comparison with the input gesture. A classification process by using Neural Networks approach is applied to investigate the individual factor that contribute to the efficiency of the recognition system. In this study, by examining the characteristics of hand trajectories, motion features are selected and classified by using a statistical approach.