RIPA: Real-Time Image Privacy Alert System


The problem of privacy and security threats arising from images uploaded onto popular social media and content sharing websites is prevalent now more than ever. As our digital footprints grow exponentially, the need to find a solution to these problems has become that much more significant. In order to address these problems, a lot of research work has been carried out for image privacy protection through privacy policy recommendations and configurations. Due to the recent advancement in the field of computer vision and deep learning we can now gain more detailed insights about the context of an image and about the relationships between objects within it, this makes it possible to better address these problems. The privacy and security threats arising from an image uploaded on-line are not only limited to the data owners. Unlike previous works that are mostly focused on individual privacy policies, we take into account privacy concerns of multiple objects depicted on the same photo (even people, animals or other objects in the background of a scenery photo) whereby these privacy concerns may not be those from the user who uploads the photo. Specifically, we first build a general knowledge base by leveraging convolution neural networks to classify sensitive and non-sensitive image content and then use our proposed metadata analysis module to analyze metadata embedded within the image. Next, we extract objects present in the photo and validate if there is any privacy violation of the objects' privacy concerns. If any sensitive object is found, we toggle the object and issue a privacy violation alert to the user who is uploading the image as well as the service provider.

Meeting Name

4th IEEE International Conference on Collaboration and Internet Computing, CIC 2018 (2018: Oct. 18-20, Philadelphia, PA)


Computer Science


This work was funded by National Science Foundation under project CNS-1651455 and CNS-1564101.

Keywords and Phrases

Convolution; Knowledge based systems; Knowledge management; Metadata; Security systems; Convolution neural network; Geolocations; Individual privacy; Metadata analysis; Privacy and security; Privacy protection; Problem of privacy; Transfer learning; Deep learning; Geo-location; Image privacy

International Standard Book Number (ISBN)


Document Type

Article - Conference proceedings

Document Version


File Type





© 2018 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.

Publication Date

01 Oct 2018