PRNU-based Image Classification of Origin Social Network with CNN

Abstract

A huge amount of images are continuously shared on social networks (SNs) daily and, in most of cases, it is very difficult to reliably establish the SN of provenance of an image when it is recovered from a hard disk, a SD card or a smartphone memory. During an investigation, it could be crucial to be able to distinguish images coming directly from a photo-camera with respect to those downloaded from a social network and possibly, in this last circumstance, determining which is the SN among a defined group. It is well known that each SN leaves peculiar traces on each content during the upload-download process; such traces can be exploited to make image classification. In this work, the idea is to use the PRNU, embedded in every acquired images, as the "carrier" of the particular SN traces which diversely modulate the PRNU. We demonstrate, in this paper, that SNmodulated noise residual can be adopted as a feature to detect the social network of origin by means of a trained convolutional neural network (CNN).

Media forensics on social media platforms: a survey

“…As it is widely known in multimedia forensics, such operations can be detected and characterized by analyzing the image signal (i.e., the values in the pixel domain or in transformed domains), where distinctive patterns can be exposed. This approach is followed in [32,[34][35][36] for platform provenance analysis, where the image signal is pre-processed to extract a feature representation (see Section 4.3).…”

“…Alternatively, the approach in [36] explores the use of the PRNU noise as a carrier of traces left by different platforms. To this purpose, a wavelet-based denoising filter is applied to each image patch to obtain a noise residual, that is then fed to the ML classifier.…”

Coherence of PRNU weighted estimations for improved source camera identification

“…In addition, it allows us to make a simple but crucial observation. Without loss of generality, we consider only the numerator in (8). Since the two terms are expected to be zero-mean, it corresponds to the inner product between K N and K d k , i.e.…”

“…Several methods have been proposed in the literature in the last years and they mainly depend on the application purpose. In fact, the working scenario can often change, requiring different operative procedures to get the final result with the expected/required accuracy [8,12,14,21,25,36,37]. The most common working scenarios are:…”

“…In the second case classification methods, mainly based on clustering techniques, are used [17,19,25]. In both cases, adopted methods are required to be robust to PRNU extraction procedures as well as data coming from different sources (notebook, social network [8], smartphone [37], etc.). As a matter of fact, in forensic investigations very few data often are available.…”