Byte-Level Object Identification for Forensic Investigation of Digital Images
Abdul Rehman Javed, Zunera Jalil
Abstract
Lately, digital data has increased a key role in providing and sharing information. Pictures and video recordings are utilized to pass on convincing messages to be utilized under a few unique situations, from propaganda to coercing. The majority of the effort in the present digital crime investigation network lies in the acquisition, retrieval, and investigation of existing data from digital machines. It is a time consuming and a humanly difficult task to collect, process and analyze each media content manually. In this paper, we provide a novel approach that solves a real-time problem for an investigator while investigating the suspect machine. Our approach acquires all image data at byte level from the suspect machine, perform fast and accurate object detection resorting to the deep learning-based algorithm and present high-level illustration of images containing suspicious object and unique objects that can be presented as evidence. Our approach aims to flag photos where suspicious objects are detected. Performance and time consumption wise, this study confirms the importance of automated object detection in digital forensics.