Spatio-temporal 3D surface matching for hand gesture recognition using ICP algorithm

摘要: With emerging new applications like virtual reality, different algorithms for human action and gesture recognition have been proposed. In this paper, a method the of moving hand gestures is presented. The proposed algorithm based on representation motion as spatio-temporal 3D surfaces. Then, surface matching used to recognize gesture. To form motion, we first apply necessary preprocessing video frames extract contours. by normalizing overlapping contours in frames, construct gesture, match input with surfaces database. For purpose, utilize ICP find compensate transformation between well similarity measure them. real-world applications, continuous results sequence disjointed gestures, which called propose an estimates probable then divides iteratively true gestures. Finally, applying robust algorithm, recognized. We tested American sign language showed rate 94 % 93.9 experimental efficiency noise well.