Forensic Source Identification of OSN Compressed Images

Source camera identification is used in legal applications involving cybercrime, terrorism, pornography, etc. It is a digital forensic way to map the image to its authentic source. In today’s digital era online social networks have become a great source of image transmission as well as mapping of the OSNs images to its source device has become difficult due to its implicit compression technique and altering of metadata. In this paper, we propose a deep learning model for SCI, on images downloaded from Facebook, and Whatsapp. We adapt the ResNet50 network and add our own layer head to fine-tune the model for the classification of source cameras of OSNs compressed images. It can be seen by observing the experimental results that the proposed technique addresses the results efficiently for images downloaded from OSNs.