K
Ke Zhang
Researcher at University of Southern California
Publications - 11
Citations - 1285
Ke Zhang is an academic researcher from University of Southern California. The author has contributed to research in topics: Automatic summarization & Supervised learning. The author has an hindex of 8, co-authored 11 publications receiving 1030 citations. Previous affiliations of Ke Zhang include Fudan University.
Papers
More filters
Posted Content
Video Summarization with Long Short-term Memory
TL;DR: Long Short-Term Memory (LSTM), a special type of recurrent neural networks are used to model the variable-range dependencies entailed in the task of video summarization to improve summarization by reducing the discrepancies in statistical properties across those datasets.
Book ChapterDOI
Video Summarization with Long Short-Term Memory
TL;DR: In this paper, the task of video summarization is cast as a structured prediction problem, and LSTM is used to model the variable-range temporal dependency among video frames to derive both representative and compact video summaries.
Proceedings ArticleDOI
Summary Transfer: Exemplar-Based Subset Selection for Video Summarization
TL;DR: In this paper, the authors propose a keyframe-based video summarization method, which leverages supervision in the form of humancreated summaries to perform automatic keyframe based summarization.
Book ChapterDOI
Retrospective Encoders for Video Summarization
Ke Zhang,Kristen Grauman,Fei Sha +2 more
TL;DR: This paper proposes to augment standard sequence learning models with an additional “retrospective encoder” that embeds the predicted summary into an abstract semantic space that outperforms existing ones by a large margin in both supervised and semi-supervised settings.
Posted Content
Summary Transfer: Exemplar-based Subset Selection for Video Summarization
TL;DR: A novel subset selection technique that leverages supervision in the form of humancreated summaries to perform automatic keyframe-based video summarization, and shows how to extend the method to exploit semantic side information about the video's category/ genre to guide the transfer process by those training videos semantically consistent with the test input.