In recent years, organizing social media by social event has drawn increasing attentions with the increasing amounts of rich-media content taken during an event. In this paper, we address the social event enrichment and summarization problem and propose a demonstration system \(E^2SGM\) to summarize the event with relevant media selected from a large-scale user contributed media dataset. In the proposed method, the relevant candidate medias are first retrieved by coarse search method. Then, a graph ranking algorithm is proposed to rank media items according to their relevance to the given event. Finally, the media items with high ranking scores are structured following a chronologically ordered layout and the textual metadata are extracted to generate the tag cloud. The work is concluded in an intuitive event summarization interface to help users grasp the essence of the event.
[1]
Benoit Huet,et al.
Heterogeneous features and model selection for event-based media classification
,
2013,
ICMR.
[2]
Mohan S. Kankanhalli,et al.
Real-life events in multimedia: detection, representation, retrieval, and applications
,
2013,
Multimedia Tools and Applications.
[3]
Vasileios Mezaris.
Social event detection at MediaEval: a 3-year retrospect of tasks and results
,
2014
.
[4]
Hung-Khoon Tan,et al.
Beyond search: Event-driven summarization for web videos
,
2011,
TOMCCAP.
[5]
Xian-Sheng Hua,et al.
Towards a Relevant and Diverse Search of Social Images
,
2010,
IEEE Transactions on Multimedia.