Postgraduate Student - PhD, University of Adelaide
Does temperature affect the fingerprint that links images to the source camera?
Each digital camera has a latent fingerprint that is embedded in each image that is taken. The fingerprint is a result of minor differences in each pixel on the image sensor Using a wavelet filter methods we are able to extract the fingerprint in each image. We are then able to link this fingerprint back to the camera fingerprint using a statistical correlation.
What is currently unknown is how temperature affects this fingerprint. This information is needed if this method is ever to be used in a court of law with any certainty for criminal investigations.
An electrical phenomenon, known as dark current, results in an image sensor outputting a low-level signal even when there is no light illuminating the sensor itself. This dark current is temperature dependent and increases exponentially as temperature increases linearly. If the dark current results in swamping our fingerprint signal then the fingerprint is not stable across temperatures and hence, may not be a reliable method for linking images to cameras for forensic purposes.
Abstract: An imaging circuit includes at least one photosensitive device that provides an output in response to at least one photon and a compensation circuit configured to provide dark current compensation for the output of said photosensitive device. The applied compensation uses temperature information and temperature dependent calibration information.
Pub.: 12 Apr '16, Pinned: 02 Aug '17
Abstract: The increased diffusion of digital images generated by mobile devices trough social networks, personal and professional communications, etc. is self-evident. This creates potential problems because some of these images may be used as supporting evidence for different criminal cases. In this paper, algorithms are proposed based on sensor noise and wavelet transforms which can alter the information which is usually employed to find the source of an image, and forge it so that it could point to a different, unrelated device. In the state of art we will show that there are already some algorithms capable of carrying out these manipulations, but they generally need much more and more complex data than our proposal. They also generally need physical access to the camera whose generated images you want to tamper. Our proposal algorithm to destruct the image identifiable data, only needs the picture which will be anonymised. Also, our proposal to forge the image identifiable data only needs a set of photos from the attacker camera, and the picture to be tampered. In particular, it does not need access to the camera that will be falsely linked to the picture. These scenarios are the most common and realistic. The algorithms proposed will help to strengthen existing techniques and develop new forensic approaches for mobile image source identification that will be more robust against attacks.
Pub.: 11 Nov '16, Pinned: 02 Aug '17
Abstract: Source camera identification is one of the emerging field in digital image forensics, which aims at identifying the source camera used for capturing the given image. The technique uses Photo Response Non-Uniformity (PRNU) noise as a camera fingerprint, as it is found to be one of the unique characteristic which is capable of distinguishing the images even if they are captured from similar cameras. Most of the existing PRNU based approaches are very sensitive to the random noise components existing in the estimated PRNU, and also they are not robust when some simple manipulations are performed on the images. Hence a new feature based approach of PRNU is proposed for the source camera identification by choosing the features which are robust for image manipulations. The PRNU noise is extracted from the images using wavelet based denoising method and is represented by Higher Order Wavelet Statistics (HOWS), which are invariant features for image manipulations and geometric variations. The features are fed to support vector machine classifiers to identify the originating source camera for the given image and the results have been verified by performing ten-fold cross validation technique. The experiments have been carried out using the images captured from various cell phone cameras and it demonstrated that the proposed algorithm is capable of identifying the source camera of the given image with good accuracy. The developed technique can be used for differentiating the images, even if they are captured from similar cameras, which belongs to same make and model. The analysis have also showed that the proposed technique remains robust even if the images are subjected to simple manipulations or geometric variations.
Pub.: 03 Nov '16, Pinned: 02 Aug '17