We present a user-based method that detects regions of interest within a video in order to provide video skims and video summaries.\r\nPrevious research in video retrieval has focused on content-based techniques, such as pattern recognition algorithms that attempt\r\nto understand the low-level features of a video.We are proposing a pulse modeling method, which makes sense of a web video by\r\nanalyzing users� Replay interactions with the video player. In particular, we have modeled the user information seeking behavior\r\nas a time series and the semantic regions as a discrete pulse of fixed width. Then, we have calculated the correlation coefficient\r\nbetween the dynamically detected pulses at the local maximums of the user activity signal and the pulse of reference. We have\r\nfound that users� Replay activity significantly matches the important segments in information-rich and visually complex videos,\r\nsuch as lecture, how-to, and documentary.The proposed signal processing of user activity is complementary to previous work in\r\ncontent-based video retrieval and provides an additional user-based dimension for modeling the semantics of a social video on the\r\nweb.
Loading....