Graduate School and Research Center in Digital Sciences

Event-based media enrichment using an adaptive probabilistic hypergraph model

Liu, Xueliang; Wang, Meng; Yin, Bao-Cai; Huet, Benoit; Li, Xuelong

IEEE Transactions on Cybernetics, December 2014, Vol. PP, N°99, ISSN: 2168-2267

Nowadays, with the continual development of digital capture technologies and social media services, a vast number of media documents are captured and shared online to help attendees record their experience during events. In this paper, we present a method combining semantic inference and multimodal analysis for automatically finding media content to illustrate events using an adaptive probabilistic hypergraph model. In this model, media items are taken as vertices in the weighted hypergraph and the task of enriching media to illustrate events is formulated as a ranking problem. In our method, each hyperedge is constructed using the K-nearest neighbors of a given media document. We also employ a probabilistic representation, which assigns each vertex to a hyperedge in a probabilistic way, to further exploit the correlation among media data. Furthermore, we optimize the hypergraph weights in a regularization framework, which is solved as a second-order cone problem. The approach is initiated by seed media and then used to rank the media documents using a transductive inference process. The results obtained from validating the approach on an event dataset collected from EventMedia demonstrate the effectiveness of the proposed approach.

Document Doi Bibtex

Title:Event-based media enrichment using an adaptive probabilistic hypergraph model
Keywords:Event enrichment, hypergraph, transductive learning.
Type:Journal
Language:English
City:
Date:
Department:Data Science
Eurecom ref:4484
Copyright: © 2014 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Bibtex: @article{EURECOM+4484, doi = {http://dx.doi.org/10.1109/TCYB.2014.2374755}, year = {2014}, month = {12}, title = {{E}vent-based media enrichment using an adaptive probabilistic hypergraph model}, author = {{L}iu, {X}ueliang and {W}ang, {M}eng and {Y}in, {B}ao-{C}ai and {H}uet, {B}enoit and {L}i, {X}uelong}, journal = {{IEEE} {T}ransactions on {C}ybernetics, {D}ecember 2014, {V}ol. {PP}, {N}°99, {ISSN}: 2168-2267}, url = {http://www.eurecom.fr/publication/4484} }
See also: