PROJECT TITLE :

Image origin classification based on social network provenance - 2017

ABSTRACT:

Recognizing data regarding the origin of a digital image has been individuated as a vital task to be tackled by the image forensic scientific community. Understanding one thing on the previous history of an image could be strategic to deal with any successive assessment to be created on it: knowing the type of device used for acquisition or, higher, the model of the camera may focus investigations during a specific direction. Sometimes simply revealing that a determined post-processing, such as an interpolation or a filtering, has been performed on a picture could be of fundamental importance to go back to its provenance. This paper locates in such a context and proposes an innovative method to inquire if a picture derives from a social network and, in specific, strive to differentiate from, which one has been downloaded. The technique relies on the belief that every social network applies a peculiar and largely unknown manipulation that, however, leaves some distinctive traces on the image; such traces can be extracted to feature every platform. By resorting at trained classifiers, the presented methodology is satisfactorily in a position to discern totally different social network origins. Experimental results dispensed on numerous image datasets and in numerous operative conditions witness that such a distinction is potential. Further, the proposed methodology is additionally ready to go back to the initial JPEG quality issue the image had before being uploaded on a social network.