Fingerprint recognition performance is affected by many factors. One of these is defective pixels caused by ageing effects of the image sensor. The authors investigate the impact of these image sensor ageing related pixel defects on the performance of different fingerprint (NBIS, VeriFinger, FingerCode and Phase Only Correlation) recognition systems. Their performances are compared against each other to quantify the differences in the impact. In practice, besides image sensor ageing related effects, other influences are also present. As the authors aim to evaluate the impact of the defective pixels only, disregarding subject ageing and other external influences, it is not possible to use real image data. Instead, an experimental study utilising an ageing simulation algorithm introducing hot and stuck pixels is conducted on the FVC2002 and FVC2004 data sets, including tests with different denoising approaches trying to mitigate the effects of image sensor ageing while maintaining the baseline recognition accuracy.

Identifying the source camera of a digital image using the photo response non-uniformity (PRNU) is known as camera identification. Since digital image sensors are widely used in biometrics, it is natural to perform this investigation with biometric sensors. In this study, the authors focus on a slightly different task, which consists in clustering images with the same source sensor in a data set possibly containing images from multiple unknown distinct biometric sensors. Previous work showed unclear results because of the low quality of the extracted PRNU. They adopt different PRNU enhancement techniques together with the generation of PRNU fingerprints from uncorrelated data in order to clarify the results. Thus they propose extensions of existing source sensor attribution techniques which make use of uncorrelated data from known sensors and apply them in conjunction with existing clustering techniques. All techniques are evaluated on simulated data sets containing images from multiple sensors. The effects of the different PRNU enhancement approaches on the clustering outcome are measured by considering the relation between cohesion and separation of the clusters. Finally, an assessment on whether the PRNU enhancement techniques have been able to improve the results is given.

The authors propose a new method for sclera quality measure and segmentation under relaxed imaging constraints. In particular, for sclera image, they propose a new quality measure approach based on a focus measure. In addition, they propose a new fusion method for sclera segmentation which uses pixel properties of both the sclera area and the skin around the eye. Furthermore, sclera template rotation alignment and distance scaling methods are proposed to minimise the error rates when noisy eye images are captured at-a-distance and on-the-move, together with overcoming head pose rotation. Then, a performance analysis on exploited eye images using the Excellent, the Good, the Bad, and the Ugly (EGBU) classification technique for image quality is used to evaluate system performance. Eye images captured under relaxed imaging constraints using four camera devices within the UBIRIS.v2 and MICHE mobile databases are utilised to evaluate the proposed sclera recognition system, with the UBIRIS.v1 database as a reference. Results in terms of sclera image quality measure and sclera segmentation are promising and describe the effect and challenges of using relaxed imaging conditions on sclera recognition system.

Substantial progress has been achieved in voice-based biometrics in recent times but a variety of challenges still remain for speech research community. One such obstacle is reliable speaker authentication from speech signals degraded by lossy compression. Compression is commonplace in modern telecommunications, such as mobile telephony, VoIP services, teleconference, voice messaging or gaming. In this study, the authors investigate the effect of lossy speech compression on text-independent speaker verification. Voice biometrics performance is evaluated on clean speech signals distorted by the state-of-the-art narrowband (NB) as well as wideband (WB) speech codecs. The tests are performed in both channel-matched and channel-mismatched scenarios. The test results show that coded WB speech improves voice authentication precision by 1–3% of equal error rate over coded NB speech, even at the lowest investigated bitrates. It is also shown that the enhanced voice services codec does not provide better results than the other codecs involved in this study.

Voice disguise by impersonation is often used in voice-based crimes by perpetrators who try to evade identification while sounding genuine. Voice evidence from these crimes is analysed to both detect impersonation, and match the impersonated voice to the natural voice of the speaker to prove its correct ownership. There are interesting situations, however, where a speaker might be confronted with voice evidence that perceptually sounds like their natural voice but may deny ownership of it, claiming instead that it is the production of an expert impersonator. This is a bizarre claim, but plausible since the human voice has a great degree of natural variation. It poses a difficult forensic problem: instead of detecting impersonation one must now prove the absence of it, and instead of matching the evidence with the natural voice of the person one must show that they cannot not have a common originator. The authors address the problem of disproving the denial of voice ownership from an articulatory-phonetic perspective, and propose a hypothesis-testing framework that may be used to solve it. The authors demonstrate their approach on data comprising voices of prominent political figures in USA, and their expert impersonators.

Parallel factor analysis 2 (PARAFAC2) is employed to reduce the dimensions of visual and aural features and provide ranking vectors. Subsequently, score level fusion is performed by applying a support vector machine (SVM) classifier to the ranking vectors derived by PARAFAC2 to make gender and age interval predictions. The aforementioned procedure is applied to the Trinity College Dublin Speaker Ageing database, which is supplemented with face images of the speakers and two single-modality benchmark datasets. Experimental results demonstrate the advantage of using combined aural and visual features for both prediction tasks.

Gait recognition systems can capture biometrical information from a distance and without the user's active cooperation, making them suitable for surveillance environments. However, there are two challenges for gait recognition that need to be solved, namely when: (i) the walking direction is unknown and/or (ii) the subject's appearance changes significantly due to different clothes being worn or items being carried. This study discusses the problem of gait recognition in unconstrained environments and proposes a new system to tackle recognition when facing the two listed challenges. The system automatically identifies the walking direction using a perceptual hash (PHash) computed over the leg region of the gait energy image (GEI) and then compares it against the PHash values of different walking directions stored in the database. Robustness against appearance changes are obtained by decomposing the GEI into sections and selecting those sections unaltered by appearance changes for comparison against a database containing GEI sections for the identified walking direction. The proposed recognition method then recognises the user using a majority decision voting. The proposed view-invariant gait recognition system is computationally inexpensive and outperforms the state-of-the-art in terms of recognition performance.

This work explores human intervention to improve Automatic Signature Verification (ASV). Significant efforts have been made in order to improve the performance of ASV algorithms over the last decades. This work analyzes how human actions can be used to complement automatic systems. Which actions to take and to what extent those actions can help state-of-the-art ASV systems is the final aim of this research line. The analysis at classification level comprises experiments with responses from 500 people based on crowdsourcing signature authentication tasks. The results allow to establish a human baseline performance and comparison with automatic systems. Intervention at feature extraction level is evaluated using a self-developed tool for the manual annotation of signature attributes inspired in Forensic Document Experts analysis. We analyze the performance of attribute-based human signature authentication and its complementarity with automatic systems. The experiments are carried out over a public database including the two most popular signature authentication scenarios based on both online (dynamic time sequences including position and pressure) and offline (static images) information. The results demonstrate the potential of human interventions at feature extraction level (by manually annotating signature attributes) and encourage to further research in its capabilities to improve the performance of ASV.

Handwriting represents an important modality in both biometric and forensic domains. while the first field appears to be a well-studied area with many approaches for acquisition, feature extraction and classification, the latter is rather novel with respect to digitised, automated methods. In prior works, authors of this article have shown that forensic handwriting traces can be processed by means of high-resolution 2D/3D surficial scanners. These devices allow the acquisition at a vertical resolution down to a few nanometres. Pre-processing methods have been identified, which allow for spatial segmentation of writing traces and for de-noising of 2D and 3D data, leading to a clear visual enhancement of the writing traces based on topography information. Further, authors have suggested a novel benchmarking method for forensic trace analysis based on feature sets originally suggested for biometric handwriting and comparing the forensic recognition performance by means of equal error rates, achieved from biometric sensors and forensic 2D/3D sensors in parallel. This article summarises and extents these works in two main aspects: on one side, it presents significantly extended experiments with respect to test set size and feature sets utilised for classification. On the other side, authors suggest new optimisation approaches to the segmentation algorithm.