An Inception-Based Data-Driven Ensemble Approach to Camera Model Identification
Anselmo Ferreira, Han Chen, Bin Li, Jiwu Huang · 2018
Proposing digital image forensic approaches for linking a suspected image to its source is of particular interest nowadays, as they can be useful to link the ownership of an investigated image with sensitive content and even help to spot image manipulations. In this paper, we propose a data-driven based solution composed of processing the output of two Convolutional Neural Networks based on the Inception philosophy. We show the effectiveness of the proposed model for the difficult problem of camera model identification, considering high-resolution images and also small patches.