On the Sensor Pattern Noise Estimation in Image Forensics: A Systematic Empirical Evaluation

Mustafa Al-Ani, Fouad Khelifi

Research output: Contribution to journalArticlepeer-review

35 Citations (Scopus)
7 Downloads (Pure)

Abstract

Extracting a fingerprint of a digital camera has fertile applications in image forensics, such as source camera identification and image authentication. In the last decade, Photo Response Non_Uniformity (PRNU) has been well established as a reliable unique fingerprint of digital imaging devices. The PRNU noise appears in every image as a very weak signal, and its reliable estimation is crucial for the success rate of the forensic application. In this paper, we present a novel methodical evaluation of 21 state-of-the-art PRNU estimation/enhancement techniques that have been proposed in the literature in various frameworks. The techniques are classified and systematically compared based on their role/stage in the PRNU estimation procedure, manifesting their intrinsic impacts. The performance of each technique is extensively demonstrated over a large-scale experiment to conclude this case-sensitive study. The experiments have been conducted on our created database and a public image database, the 'Dresden image database
Original languageEnglish
Pages (from-to)1067-1081
JournalIEEE Transactions on Information Forensics and Security
Volume12
Issue number2
DOIs
Publication statusPublished - 15 Dec 2016

Fingerprint

Dive into the research topics of 'On the Sensor Pattern Noise Estimation in Image Forensics: A Systematic Empirical Evaluation'. Together they form a unique fingerprint.

Cite this