Hand Gesture Recognition by using Logical Heuristics

概要

論文の詳細を見る
We present a method to recognize hand gestures, based on skin color detection and heuristic conditions using convex hull and convexity defects. In this research, skin color is detected in the preliminary task, for which we propose an algorithm to obtain a smoother skin pixel binary mask in order to recognize a hand in various complex backgrounds by using adaptive skin color scheme. Gesture recognition is one of the essential techniques to build user friendly interfaces. As it is costly to prepare large gesture database for training or manufacture data gloves for gesture recognition. We try to find the best heuristic conditions which work well for both video and still images taken from a standard digital camera. Our method is able to recognize six different hand gestures in complex backgrounds and can be applied to various human-computer interaction interfaces. Our results showed that the heuristic conditions performed well in recognizing gestures of a hand with good precision and applicability.
2012-03-14