Photo response non-uniformity (PRNU) noise is a sensor pattern noise characterizing imperfections in the imaging device. The PRNU is a unique noise for each sensor device, and it has been generally utilized in the literature for source camera identification and image authentication. In video forensics, the traditional approach estimates the PRNU by averaging a set of residual signals obtained from multiple video frames. However, due to lossy compression and other non-unique content-dependent noise components that interfere with the video data, constant averaging does not take into account the intensity of these undesirable noise components which are content-dependent. Different from the traditional approach, we propose a video PRNU estimation method based on weighted averaging. The noise residual is first extracted for each single video. Then, the estimated noise residuals are fed into a weighted averaging method to optimize PRNU estimation. Experimental results on two video datasets captured by various smartphone devices have shown a significant gain obtained with the proposed approach over the conventional state-of-the-art one.

This paper introduces a standard digital picture dataset specifically designed for temporal digital image forensics. The database, called Northumbria Temporal Image Forensics (NTIF), consists of natural images with full high resolution of indoor and outdoor scenes. The images are organized in temporal order with regular acquisition timeslots spanned over for 94 weeks using ten digital camera devices. 41,684 images were captured from 10 digital cameras belonged to different models and brands. To this end, the subset of images has been annotated with labels spanning over categories based on the temporal factor of one to two weeks. Constructing such a large-scale temporal image database has been a challenging and enduring process. During the construction of NTIF, ethics were fully considered. The proposed dataset will be freely accessible to benefit all researchers in image forensics from academia and industry. This paper aims to describe the NTIF database and highlight the changes in Sensor Pattern Noise over time. Experiments have been conducted in which the correlations between noise residuals appear to be sensitive to the acquisition time of the respective digital images. The results show a clearly different pattern of correlations when the images are captured in different timeslots as compared to those images acquired within the same timeslots.

Photo response non-uniformity (PRNU) noise is a sensor pattern noise characterizing the imaging device. It has been broadly used in the literature for source camera identification and image authentication. The abundant information that the sensor pattern noise carries in terms of the frequency content makes it unique, and hence suitable for identifying the source camera and detecting image forgeries. However, the PRNU extraction process is inevitably faced with the presence of image-dependent information as well as other non-unique noise components. To reduce such undesirable effects, researchers have developed a number of techniques in different stages of the process, i.e., the filtering stage, the estimation stage, and the post-estimation stage. In this paper, we present a new PRNU-based source camera identification and verification system and propose enhancements in different stages. First, an improved version of the locally adaptive discrete cosine transform filter is proposed in the filtering stage. In the estimation stage, a new weighted averaging technique is presented. The post-estimation stage consists of concatenating the PRNUs estimated from color planes in order to exploit the presence of physical PRNU components in different channels. Experimental results on two image data sets acquired by various camera devices have shown a significant gain obtained with the proposed enhancements in each stage as well as the superiority of the overall system over related state-of-the-art systems.

Photo-response non-uniformity (PRNU) noise has been well established as a reliable fingerprint of imaging sensors for source camera identification and other forensic applications. In this paper, we introduce a novel denoising method for PRNU noise extraction that considerably outperforms the existing techniques. The rationale is to use as little as one adjacent pixel in spatial domain filtering to suppress the pixel-to-pixel correlation in the estimation of the (supposedly white) PRNU noise. Experimental results are presented based on an image database of ten cameras of different models and makes. These results are embodied in the receiver operating characteristic (ROC) of source camera identification.

Sensor pattern noise has been broadly used in the literature for image authentication and source camera identification. The abundant information that a sensor pattern noise carries in terms of the frequency content makes it unique and hence suitable for source camera identification. The traditional approach for estimating the sensor pattern noise uses a set of images to estimate a pattern residual signal from each image. The estimated residual signals are then averaged to obtain the sensor pattern noise. This is based on the assumption that each residual signal is a noisy observation of the sensor pattern noise. Such an assumption is well justified in practice because the images are acquired under different conditions making the corresponding residual signals distinct from each other. For instance bright images provide better sensor pattern noise estimation than dark images. Also, saturated pixels cause undesirable noise in residual signals. Inspired by this observation, a weighted averaging approach is proposed for efficient sensor pattern noise estimation. The proposed approach has been validated with two sensor pattern noise estimation techniques from the literature and significant improvements have been shown through experimental results.