Motion Panorama

Abstract

Access to multimedia data and multimedia services is becoming increasingly common in networked mobile environments comprising of clients using mobile handheld devices, such as personal digital assistants (PDAs), iPads and cellular devices. One of the natural limitations of typical handheld mobile devices is that they are resource constrained, i.e., constrained by their battery power capacity, viewing time limit and in many situations, by the available network bandwidth connecting them with video servers. A typical video server is also resource constrained in terms of its computing power and storage capacity and has to ensure efficient utilization of its resources in order to provide an optimal quality of service to its clients with minimal client-experienced latency. Thus, in mobile multimedia environments, the original video content often needs to be “transcoded” in order to fulfill the clients’ requests while satisfying simultaneously the various client-side and system-wide resource constraints. A motion panorama is an efficient and compact “transcoded” representation of the underlying video. In the project, sequential motion panorama algorithm is implemented and tested using video data in Matlab. Faculty Adviser: Yong Wei