Sökning: onr:"swepub:oai:DiVA.org:umu-175288" >
Privacy-Preserving ...
Privacy-Preserving Visual Content Tagging using Graph Transformer Networks
-
- Vu, Xuan-Son (författare)
- Umeå universitet,Institutionen för datavetenskap,Deep Data Mining Group
-
- Le, Duc-Trong (författare)
- University of Engineering and Technology, VNU, Vietnam
-
Edlund, Christoffer (författare)
-
visa fler...
-
- Lili, Jiang (författare)
- Umeå universitet,Institutionen för datavetenskap,Deep Data Mining Group
-
- Nguyen, D. Hoang (författare)
- School of Computing Science, University of Glasgow, Singapore
-
visa färre...
-
(creator_code:org_t)
- 2020-10-12
- 2020
- Engelska.
-
Ingår i: Proceedings of the 28th ACM International Conference on Multimedia (MM ’20). - New York, NY, USA : ACM Digital Library. - 9781450379885 ; , s. 2299-2307
- Relaterad länk:
-
https://doi.org/10.1...
-
visa fler...
-
https://dl.acm.org/d...
-
https://urn.kb.se/re...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- With the rapid growth of Internet media, content tagging has become an important topic with many multimedia understanding applications, including efficient organisation and search. Nevertheless, existing visual tagging approaches are susceptible to inherent privacy risks in which private information may be exposed unintentionally. The use of anonymisation and privacy-protection methods is desirable, but with the expense of task performance. Therefore, this paper proposes an end-to-endframework (SGTN) using Graph Transformer and Convolutional Networks to significantly improve classification and privacy preservation of visual data. Especially, weemploy several mechanisms such as differential privacy based graph construction and noise-induced graph transformation to protect the privacy of knowledge graphs. Our approach unveils new state-of-the-art on MS-COCO dataset in various semi-supervised settings. In addition, we showcase a real experiment in the education domain to address the automation of sensitive document tagging. Experimental results show that our approach achieves an excellent balance of model accuracy and privacy preservation on both public and private datasets.
Ämnesord
- NATURVETENSKAP -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Computer Sciences (hsv//eng)
Publikations- och innehållstyp
- ref (ämneskategori)
- kon (ämneskategori)
Hitta via bibliotek
Till lärosätets databas