On the repudiability of device identification and image integrity verification using sensor pattern noise

Abstract

In this work we study the power of the methods for digital device identification and image integrity verification, which rely on sensor pattern noise as device signatures, and the repudiability of the conclusions drawn from the information produced by this type of methods. We prove that the sensor pattern noise existing in the original images can be destroyed so as to confuse the investigators. We also prove that sensor pattern noise of device A can be easily embedded in the images produced by another device B so that the device identifier would mistakenly suggest that the images were produced by device A, rather than by B, and mislead forensic investigations. © Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering 2010.