An analysis of optical contributions to a photo-sensor's ballistic fingerprints

Abstract

Lens aberrations have previously been used to determine the provenance of an image. However, this is not necessarily unique to an image sensor, as lens systems are often interchanged. Photo-response non-uniformity noise was proposed in 2005 by Lukáš, Goljan and Fridrich as a stochastic signal which describes a sensor uniquely, akin to a "ballistic" fingerprint. This method, however, did not account for additional sources of bias such as lens artefacts and temperature.In this paper, we propose a new additive signal model to account for artefacts previously thought to have been isolated from the ballistic fingerprint. Our proposed model separates sensor level artefacts from the lens optical system and thus accounts for lens aberrations previously thought to be filtered out. Specifically, we apply standard image processing theory, an understanding of frequency properties relating to the physics of light and temperature response of sensor dark current to classify artefacts. This model enables us to isolate and account for bias from the lens optical system and temperature within the current model.

Reverse engineering the Raspberry Pi Camera V2: A study of pixel non-uniformity using a scanning electron microscope

“…We have previously shown the contamination of lens effects within the SPN methodology [7,8,9]. In [10] the non-uniform output of a photodiode was presented.…”

Identifying the Digital Camera from Natural Images Using Residual Noise and the Jensen–Shannon Divergence

“…Note that it includes a Wiener filter to remove artifacts due to JPEG compression. It is worth noting that although the work published in 2019 by Matthews et al [36] considers other elements (lens artifacts or those produced by temperature) that could influence the identification process of source digital camera, similar to Goljan et al [21], those elements were not considered. is decision was based in the fact that in 2014, Gisolf et al [37] assured that the method proposed by Goljan et al [21] was computationally slow compared with their method, but conversely, it was more accurate.…”

Interpol review of imaging and video 2016–2019

“…Identification of images that might have a common source can also be helpful in these investigations. The developments that have been started in the period of the previous review have not been stopped and have lead to a number of new methods and software packages [ [51] , [52] , [53] , [55] , [56] , [57] , [58] , [59] , [60] , [61] , [62] , [63] , [64] , [65] , [66] , [67] , [68] , [69] , [70] , [71] , [72] , [73] , [74] , [75] , [76] , [b] , [77] , [a] , [b] , [78] , [a] , [79] , [80] , [81] , [82] , [83] , [84] , [85] , [86] , [87] , [88] , [89] , [9] ]. The most used method is based on the estimation of a specific type of fixed pattern noise in an image that is caused by PRNU - Photo Response Non Uniformity .…”

“…the frames in a video file [ [49] , [50] , [58] , [91] , [92] , [93] , [94] , [95] , [97] , [98] ], much better estimations of the PRNU pattern can be obtained by averaging techniques. In the newer cameras one has to compensate for motion compensation [ 82 , 84 , 88 , 90 ]. However several methods are presented to improve the calculation speed as well as clustering images if the camera is not available.…”