作者: James M. Rehg , Takeo Kanade
DOI: 10.1007/BFB0028333
关键词: Grayscale 、 Kinematic chain 、 Motion (physics) 、 Computer science 、 Tracking (particle physics) 、 Computer vision 、 Tracking system 、 Artificial intelligence 、 Eye tracking 、 Finger tracking
摘要: Passive sensing of human hand and limb motion is important for a wide range applications from human-computer interaction to athletic performance measurement. High degree freedom articulated mechanisms like the are difficult track because their large state space complex image appearance. This article describes model-based tracking system, called DigitEyes, that can recover 27 DOF model ordinary gray scale images at speeds up 10 Hz.