Biometrics

In this work, we study the applicability of PRNU-based sensor identification methods for finger vein imagery. We also investigate the effect of different image regions on the identification performance by looking at five different croppings with different sizes. The proposed method is tested on eight publicly available finger vein datasets. For each finger vein sensor a noise reference pattern is generated and subsequently matched with noise residuals extracted from previously unseen finger vein images. Although the final result strongly encourages the use of PRNU-based approaches for sensor identification, it can also be observed that the choice of image region for PRNU extraction is crucial. The result clearly shows that regions containing biometric trait (varying content) should be preferred over background regions containing non-biometric trait (identical content).

We examine the effectiveness of the Photo Response Non-Uniformity (PRNU) in the context of sensor identification for finger vein imagery. Experiments are conducted on eight publicly-available finger vein datasets. We apply a Wiener Filter (WF) in the frequency domain to enhance the quality of PRNU estimation and noise residual, respectively, and we use two metrics to rank PRNU similarity, i.e. Peak-to-Energy (PCE) and Normalized Cross Correlation (NCC). In the experiments, we include a dataset consisting of both real finger vein data and captured artifacts produced to assess presentation attacks. We investigate the impact of this situation on sensor identification accuracy and also try to discriminate spoofed images from non-spoof images varying decision thresholds. Results of sensor identification for finger vein imagery is encouraging, the obtained scores for classification accuracies are between 97% to 98% for different settings. Interestingly, selecting particular decision thresholds, it is also possible to discriminate real data from artificial data as used in presentation attacks.

In this work, we evaluated the effectiveness of the Photo Response Non-Uniformity (PRNU) to detect presentation/spoofing attacks for finger vein imagery. The performance is evaluated on two publicly-available finger vein presentation/spoofing attack datasets (IDIAP and SCUT-FVD). Maximum likelihood estimation (MLE) is used to estimate the sensor’s PRNU. To decide whether a query image is real or spoofed, we compare its residual to the estimated sensor PRNU using PCE and NCC as similarity measures. We observe that the classification performance is heavily dependent on the set of images used for PRNU estimation. We assume different degrees of variability in image content caused by distinct light scattering properties in real tissue and artifacts to be one of the main reasons for the differences in classification performance.

Identifying the origin of a sample image in biometric systems can be beneficial for data authentication in case of attacks against the system and initiating sensor-specific processing pipelines in sensor-heterogeneous environments. Motivated by shortcomings of the photo response non-uniformity (PRNU) based method in the biometric context, we employ eight texture classification approaches, including frequency-, spatial-, and wavelet-based methods to detect finger vein samples images’ origin. Besides, We use eight publicly available finger vein datasets and applying all eight novel classical texture descriptors and SVM classification in the suggested pipeline. A novel wavelet-based approach termed WMV demonstrated an excellent result for raw finger vein samples and the more challenging region of interest data among mentioned employed methods to identify sensor model. The observed results establish texture descriptors as effective competitors to PRNU in finger vein sensor model identification.

We study the finger vein (FV) sensor model identification task using a deep learning approach. So far, for this biometric modality, only correlation-based PRNU and texture descriptor-based methods have been applied. We employ five prominent CNN architectures covering a wide range of CNN family models, including VGG16, ResNet, and the Xception model. In addition, a novel architecture termed FV2021 is proposed in this work, which excels by its compactness and a low number of parameters to be trained. Original samples, as well as the region of interest data from eight publicly accessible FV datasets, are used in experimentation. An excellent sensor identification AUC-ROC score of 1.0 for patches of uncropped samples and 0.9997 for ROI samples have been achieved. The comparison with former methods shows that the CNN-based approach is superior and improved the results.