Geographical position spoofing detection based on camera sensor fingerprint

Abstract

Currently, position check-in on mobile devices has become a fashionable social activity. Meanwhile, criminals probably tamper the geographical position (geo-position) information to provide an alibi. Therefore, it is of importance to identify the authenticity of geo-position. To our knowledge, many current methods for geo-position spoofing detection mainly rely on geo-position information in the database. However, these methods possibly fail in the case of missing prior information or lacking rich training samples. To address that challenge, this paper proposes an alternative manner for detecting the geo-position spoofing via camera sensor fingerprint. In particular, the camera sensor fingerprint is first extracted through the images posted by an inquiry user based on the well-designed denoising filter. Second, the authenticity of the geo-position is verified by comparing the consistency of the residual noise from newly-posted images with position check-in and the unique camera sensor fingerprint from an inquiry user. Finally, the extensive experiments are conducted on the image database, that empirically indicates the relevance of our proposed simple but effective method.