Facenet: Tracking People and Acquiring Canonical Face Images in a Wireless Camera Sensor Network

We describe a method for tracking people in 2D world coordinates and acquiring canonical frontal face images that fits the sensor network paradigm. Frontal face images are particularly desireable features for tracking and identity management because they are largely invariant to day-to-day changes in appearance. This approach has been implemented and evaluated on a prototype wired camera network called FaceNet. Our primary contribution is to show how sensing the trajectories of moving objects can be exploited to acquire high quality canonical views while conserving node energy. We present an evaluation of the approach and demonstrate the tasking algorithm in action on data acquired from the FaceNet camera network.