Performance analysis of RTEPI method for real time hand gesture recognition

Abstract

Hand gestures provide a natural and intuitive way to interact with the computers, mobile devices, etc. Many researchers are working on gesture recognition where the aim is to identify and distinguish human gestures and then identified gestures are used to control applications in specific domains. Recognizing real time hand gestures is very challenging and difficult task. This paper presents a novel Real Time End Point Identification Method from a real time video of hand gesture using computer vision and Image processing techniques. The Real Time End Point Identification (RTEPI) method discussed in [1] is based on Accurate End Point Identification [2], which enables the AEPI method to work on real time hand gestures captured through web camera or laptop camera. The RTEPI method provides the correct frame for real time processing to AEPI method which then detects the accurate method. The AEPI method has been implemented to address the problems of varying background, luminance, blurring etc. Five different phases of AEPI method includes preprocessing, centroid detection, removal of unwanted objects, thinning and recognition which are already discussed in [2,3]. The paper presents the result and performance analysis of RTEPI method for all possible input patterns of real time hand gesture recognition.