作者: K. Kwon , H. Zhang , F. Dornaika
DOI: 10.1109/ROBOT.2001.932773
关键词: Animation 、 Application software 、 Pose 、 Engineering 、 Video camera 、 Iterative method 、 Computer vision 、 Artificial intelligence 、 3D pose estimation 、 Motion estimation 、 Stereopsis
摘要: We present a system to determine human hand pose from single monochrome image in real-time. formulate the problem as determining position and orientation of palm well joint angles fingers. The 3D each angle are computed simultaneously using feature points on hand. exploit an existing fast object algorithm developed by DeMenthon Davis (1995) together with heuristic minimization 1D function. make use special glove remove noise caused deformation help find correspondence between their images. demonstrate our mock-up cardboard one finger. Experimental results real wearing presented.