|
Juan Wachs, Department of Computer Science, NPS
1500-1550 2/28/2008, Glasgow East-117
Abstract: Hand gesture control systems require high learnability, usability, ergonomic design and comfort. Unfortunately, most gesture interfaces are designed with recognition accuracy as the central focus. The selection of hand gestures that consider recognition accuracy as well as the ease of learning, lack of stress, cognitively natural, and ease of implementation is still an open research question.
This work presents an analytical approach to the design of a gesture vocabulary (GV) using multiobjectives for psycho-physiological and gesture recognition factors. A meta-heuristic approach is taken by decomposing the problem into two sub-problems: (i) finding the subsets of gestures that meet a minimal accuracy requirement, and (ii) matching gestures to commands to maximize the human factors objective. The result is a set of solutions from which a Pareto optimal subset is selected. An example solution from the Pareto set is exhibited using prioritized objectives.
Speaker Bio: Juan Wachs received his B.Ed.Tech degree in Electrical Education from the ORT Academic College in Jerusalem, Israel, in 1995. He received M.Sc. in Information Systems, and a PhD degree in Intelligent Systems at Ben-Gurion University of the Negev in Israel, in 2007. Dr. Wachs is currently a postdoc from the NRC Research Associateship Program, working with Mathias Kolsch on topics related to computer vision, human-robot interfaces, multimodal interaction and gesture recognition.
View the flyer here.
|
