Venkata Udaya Sameer, B (2020) Mitigating Challenges in Image Source Attribution through Digital Forensics. PhD thesis.
|PDF (Restricted upto 26/02/2023)
The study, analysis and investigation of digital evidences in relation to cybercrime investigation, constitutes the branch of science known as digital forensics. In today’s forensic world, images collected from electronic devices related to a crime scene are proven to be extremely useful elements in forensic investigations, and to trace perpetrators related to crime scenes. Digital images play a major role in forensic investigations today, and act as the primary elements to establish legal evidences. Mapping a contentious image correctly to its source of origin, hence attribution of an image collected from crime scene to a suspect’s device, is a crucial aspect of digital forensic investigation.
In this research, we focus on the problem of source camera identification in digital image forensics. In the forensic image source identification problem, the task at hand is to associate an image under question to a suspect’s camera, thus incriminating the suspect correctly. This is known as image source verification which answers the question of whether a particular camera device has indeed captured the query image. However, practically it may not always be feasible to obtain physical access to the devices owned by every suspect. In such cases, the traditional source verification mechanisms fail, and it becomes even more challenging to identify the correct image source (device make and model). In this thesis, we address the major present-day challenges associated with forensic source camera identification, aimed towards solving the practical problems encountered by a forensic analyst during image source attribution.
In this thesis, we also investigate the biggest threat to state-of-the-art forensic source attribution techniques, which is constituted of counter-forensic attacks on digital images. We propose efficient measures to distinctly identify counter-forensic images, as well as the class of attack that they have undergone, while accurately mapping such images back to their correct sources.
|Camera; Classification; Counter–Forensics; Deep Learning; Digital Forensics; Few–Shot; Photo Response Non Uniformity (PRNU); Source Identification; Unknown–Models
|Engineering and Technology > Computer and Information Science
Engineering and Technology > Computer and Information Science > Image Processing
|Engineering and Technology > Department of Computer Science Engineering
|IR Staff BPCL
|26 Feb 2021 11:19
|15 Mar 2023 11:44
Repository Staff Only: item control page