Geometric-Imprints: A Significant Points Extraction Method for the Scan&Track Virtual Environment

are analyzed for extrapolating the 3D information inher­ ent in the scene, the well known correspondence problem We have implemented a new algorithm, called the is ever-present. Magnetic sensors, placed on several places geometric-imprint.s method for extracting key-features of on the body of the human participant, can be used for de­ a given contour. Series of points on the curve are used tecting the motion of the participant. The motion is then to find a geometric-imprint set of the contour. The algo­ used to control a synthetic human-form or an avatar of the rithm is based upon the assumption that an articulated participant. Some solutions to drive an avatar using min­ human and other figures have mostly cylindrical body­ imal sensors have been already developed [5, 6, i, 8]. In­ parts. The algorithm recursively partitions the contour teractive video-environments and wearable computers are based upon the cylindrical nature of human-silhouettes, also the emphasis in [9]. When we use encumbering de­ and a starting point. Similar topological contours are ex­ vices on the participant, we actually avoid much of the pected to produce similar geometric-imprints. Thus the correspondence problem. The correspondence problem is correspondence problem is somewhat simplified. Results severe in camera-based applications because multiple cam­ of our implementation are presented. era images are analyzed and enough information has to be extracted from these images so that the position, motion,