Ping-Hao Wu (University of Southern California), “Efficient Management Techniques for Large Scale Video” (2010)

Ping-Hao Wu (University of Southern California), “Efficient Management Techniques for Large Scale Video” Advisor: C.-C. Jay Kuo (2010)

In this dissertation, two techniques regarding the management of large video collections are investigated. First, a fast and robust duplicate video detection system is proposed. It matches video clips according to their temporal structures, which can be represented by frames that mark important events. Camera transitional behaviors are used to indicate such events. Then, an efficient signature matching algorithm based on suffix array is adopted. Experimental results show that the proposed framework is both efficient and effective in identifying duplicates. In the second topic, the focus is on the relationship between video genres and the camera usage.  Several features are designed to capture the different properties of professional and amateur video.  These features address various differences in video quality and editing effects.  Experimental results demonstrate that the camera usage can be inferred from the proposed features and, thus, reliable separation of professional and amateur video contents can be achieved.

For details, click here or contact the author.