Home // ICIW 2023, The Eighteenth International Conference on Internet and Web Applications and Services // View article
Authors:
Sergej Schultenkämper
Frederik Simon Bäumer
Keywords: Computer Vision; Privacy; Social Networks.
Abstract:
Threats to user privacy in Web 2.0 are manifold. They can arise, for example, from texts, geoinformation, images, videos or combinations of these. In order to warn users of possible threats, it is necessary to find as much relevant information as possible. However, finding and aggregating relevant, user-specific information across web platforms, such as social networks, is challenging -- not only because of the overwhelming amount of data but also due to the data quality and the great number of possible variants. In this paper, we investigate whether vision-language understanding techniques can be used to identify relevant image data and reliably extract sensitive information from these images. Our findings show that these methods are suitable for the pre-selection of relevant images, yet there are weaknesses in the extraction of information.
Pages: 1 to 6
Copyright: Copyright (c) IARIA, 2023
Publication date: June 26, 2023
Published in: conference
ISSN: 2308-3972
ISBN: 978-1-68558-069-8
Location: Nice, France
Dates: from June 26, 2023 to June 30, 2023