Multimedia content created by users (in contrast to content created by professional studios) has been growing rapidly in recent years. Most of the content uses materials from the creators' personal collections, generated by the users, or downloaded from the Internet. It is difficult to create content that requires materials from far-away locations, long-term observations, or many sources. This paper presents a system using globally deployed network cameras to retrieve images or videos for creating personalized multimedia contents. The system currently has over 60,000 cameras across the world. Through the system, users can select some of the cameras and record data at desired time, frame rate, and duration. A demonstration creates a personalized news report for the Macy's Thanksgiving Parade in New York without actually visiting New York.
Download Full PDF Version (Non-Commercial Use)