Fred W. SmithHassan MostafaviChristopher Albrecht
Video cameras are often used to record the motion of test objects. Manual quantitative measurement of the objects' location and orientation generally requires time consuming analysis of large numbers of images. This paper presents a video tracker using digital imaging techniques, that can automatically track operator-designated features in a sequence of video images. Either the centroid or the edge of a feature can be tracked. The direction of the edge is selected by the operator during track set up, and can be an any angle. Multiple objects in one image and/or multiple features on one object can be tracked. Tracking rates up to 15 images per second are possible. A human operator starts the tracking by designating the feature to be tracked, monitors the automatic tracking, and assists the tracker if necessary. An adaptive threshold maintains automated tracking through changing lighting and background conditions.
Stefan HolzreiterJohann KastnerPhilipp Wagner
Philip AlperLynn RosenfeldDaniel Kliger