Xufeng Lin

Preprocessing Reference Sensor Pattern Noise via Spectrum Equalization

Although sensor pattern noise (SPN) has been proved to be an effective means to uniquely identify digital cameras, some non-unique artifacts, shared among cameras undergo the same or similar in-camera processing procedures, often give rise to false identifications. Therefore, it is desirable and necessary to suppress these unwanted artifacts so as to improve the accuracy and reliability. In this paper, we propose a novel preprocessing approach for attenuating the influence of the non-unique artifacts on the reference SPN to reduce the false identification rate. Specifically, we equalize the magnitude spectrum of the reference SPN through detecting and suppressing the peaks according to the local characteristics, aiming at removing the interfering periodic artifacts. Combined with six SPN extractions or enhancement methods, our proposed spectrum equalization algorithm is evaluated on the Dresden image database as well as our own database, and compared with the state-of-the-art preprocessing schemes. The experimental results indicate that the proposed procedure outperforms, or at least performs comparable with, the existing methods in terms of the overall receiver operating characteristic curves and kappa statistic computed from a confusion matrix, and tends to be more resistant to JPEG compression for medium and small image blocks.

Enhancing Sensor Pattern Noise via Filtering Distortion Removal

In this work, we propose a method to obtain higher quality sensor pattern noise (SPN) for identifying source cameras. We believe that some components of SPN have been severely contaminated by the errors introduced by denoising filters and the quality of SPN can be improved by abandoning those components. In our proposed method, some coefficients with higher denoising errors are abandoned in the wavelet representation of SPN and the remaining wavelet coefficients are further enhanced to suppress the scene details in the SPN. These two steps aim to provide better SPN with higher signal-to-noise ratio (SNR) and therefore improve the identification performance. The experimental results on 2,000 images captured by 10 cameras (each responsible for 200 images), show that our method achieves better receiver operating characteristic (ROC) performance when compared with some state-of-the-art methods.

Exposing image forgery through the detection of contrast enhancement

In this paper, a novel forensic method of exposing cut-and-paste image forgery through detecting contrast enhancement is proposed. We reveal the inter-channel correlation introduced by color image interpolation, and show how a linear or nonlinear contrast enhancement can disturb this natural inter-channel dependency. We then construct a metric to measure these correlations, which are useful in distinguishing the original and contrast enhanced images. The effectiveness of the proposed algorithm is experimentally validated on natural color images captured by commercial cameras. Finally, its robustness against some anti-forensic algorithms is also discussed.

Two improved forensic methods of detecting contrast enhancement in digital images

Contrast enhancements, such as histogram equalization or gamma correction, are widely used by malicious attackers to conceal the cut-and-paste trails in doctored images. Therefore, detecting the traces left by contrast enhancements can be an effective way of exposing cut-and-paste image forgery. In this work, two improved forensic methods of detecting contrast enhancement in digital images are put forward. More specifically, the first method uses a quadratic weighting function rather than a simple cut-off frequency to measure the histogram distortion introduced by contrast enhancements, meanwhile the averaged high-frequency energy measure of his- togram is replaced by the ratio taken up by the high-frequency components in the histogram spectrum. While the second improvement is achieved by applying a linear-threshold strategy to get around the sensitivity of threshold selection. Compared with their original counterparts, these two methods both achieve better performance in terms of ROC curves and real-world cut-and-paste image forgeries. The effectiveness and improvement of the two proposed algorithms are experimentally validated on natural color images captured by commercial camera.

An Improved Algorithm for Camera Model Identification Using Inter-channel Demosaicking Traces

Most CFA (color filter array) interpolation-based digital image forensic methods characterize inter-pixel relationship with a linear model and use the estimated interpolation coefficients as features for image source camera identification. However, various CFA models and interpolation algorithms must be tried for coefficient estimation during the detection process in that the CFA pattern of an image is often unknown at the receivers end. This incurs high computational complexity. Instead of using inter-pixel correlations, Ho et al. proposed to use inter-channel demosaicking/color interpolation traces for identifying the source camera model of a test image. In this work, we propose an improved algorithm. We first extract two variance maps by estimating the variances of each component of the green-to-red and green-to-blue spectrum differences, respectively, and then take the shape and texture features of these two maps for camera model identification. Experimental results show that our method achieves better overall performance.

Source Camera Identification Issues: Forensic Features Selection and Robustness

Statistical image features play an important role in forensic identification. Current source camera identification schemes select image features mainly based on classification accuracy and computational efficiency. For forensic investigation purposes; however, these selection criteria are not enough. Consider most real-world photos may have undergone common image processing due to various reasons, source camera classifiers must have the capability to deal with those processed photos. In this work, the authors first build a sample camera classifier using a combination of popular image features, and then reveal its deficiency. Based on the experiments, suggestions for the design of robust camera classifiers are given. Copyright

Refining PRNU-based detection of image forgeries

Photo Response Non-Uniformity (PRNU) noise can be considered as a spread-spectrum watermark embedded in every image taken by the source imaging device. It has been effectively used for localizing the forgeries in digital images. The noise residual extracted from the image in question is compared with the reference PRNU in a sliding-window based manner. If their normalized cross correlation, which servers as a decision statistic, is below a pre-determined threshold (e.g., by Neyman-Pearson criterion), the center pixel in the window is declared as forged. However, the decision statistic is calculated over the forged and the non-forged regions when the sliding window falls near the boundary of the two different regions. As a result, the corresponding pixels of the forged region are probably wrongly identified as genuine ones. To alleviate this problem, we analyze the correlation distribution in the problematic region and refine the detection by weighting the decision threshold based on the altered correlation distribution. The effectiveness of the proposed refining algorithm is confirmed through the results of detecting three different kinds of realistic image forgeries.