Paper Title
Visual Tracking Methods For Mobile Devices
Abstract
Visual tracking is a time consuming process; therefore, we can hardly achieve quality tracking performance by using
mobile devices. Applications of visual tracking include object tracking and displacement estimation. In this study, we develop
a haar-like feature based method for face tracking and a 2D logarithmic search method for motion estimation. We use these
two methods alternatively for mobile display stabilization and compare the resulted performances respectively. The workflow
of the stabilization system consists of three pipelined units, i.e., tracking the motion of the mobile device, predicting the
displacement of the device, and adjusting the position of the display area to compensate for vibrations. We validated the
performance of these two methods by the upper bounds of the frame rates and the prediction errors. Experimental results
reveal that the 2D logarithmic search method outperformed the other; implying that we may not need to track objects for some
specific mobile applications, such as device moving prediction and display stabilization.
Index Terms- visual tracking, haar-like feature, 2D logarithmic search, mobile display