Image forensics of digital cameras by analysing image variations using Statistical Process Control

In this paper, we propose the novel use of Statistical Process Control (SPC) as a tool for identifying anomalies in digital cameras, by analysing image variations. We apply SPC on six low-end cameras, and one mid-range camera. By plotting the mean distribution of image pixel data through control charts, we have been able to identify a clear distinction between images captured from low-end cameras (where the variation is approximately 21%), and mid-range cameras (where the variation is approximately 1%). Control charts are used to highlight out-of-control values within the image data. By examining the data of any out-of-control images, it is possible to deduce the cause of inconsistency in the device's image acquisition process. This could ultimately lead to the identification of a stochastic feature for images taken by certain types of digital cameras, to benefit research for camera identification. In this paper, the SPC model is described in relation to its potential use in the field of image forensics for aiding camera identification, and some experimental results are shown from the six test cameras to demonstrate the effectiveness of the model.