Picture-to-Identity linking of social network accounts based on Sensor Pattern Noise

The widespread diffusion of digital imaging devices fuelled a growing interest on photo sharing through social networks. Nowadays, Internet users continuously leave visual “traces” of their presence and life on the Internet, which can constitute precious data for forensic investigators. Digital Image Forensics tools are used to analyse such images and collect evidences. One of such tools is the Sensor Pattern Noise (SPN), that is, an unique “fingerprint” left on a picture by the source camera sensor. In this paper, we propose and experimentally test a novel usage of SPN, to find social network accounts belonging to a person of interest, who has shot a given photo. We name this task Picture-to-Identity linking, and believe it can be useful in a variety of forensic cases, e.g., finding stolen camera devices, cyber-bullying, or on-line child abuse. We evaluate two methods for Picture-to-Identity linking based on two existing SPN comparison techniques, on a benchmark data set of publicly accessible social network accounts collected from the Internet. The reported results are promising and show that such technique has a practical value for forensic practitioners.