A Unified Model for Fingerprint Authentication and Presentation Attack Detection

Typical fingerprint recognition systems are comprised of a spoof detection module and a subsequent recognition module, running one after the other. In this study, we reformulate the workings of a typical fingerprint recognition system. We show that both spoof detection and fingerprint recognition are correlated tasks, and therefore, rather than performing these two tasks separately, a joint model capable of performing both spoof detection and matching simultaneously can be used without compromising the accuracy of either task. We demonstrate the capability of our joint model to obtain an authentication accuracy (1:1 matching) of TAR \(=\) 100% @ FAR \(=\) 0.1% on the FVC 2006 DB2A dataset while achieving a spoof detection ACE of 1.44% on the LivDet 2015 dataset, both maintaining the performance of stand-alone methods. In practice, this reduces the time and memory requirements of the fingerprint recognition system by 50 and 40%, respectively; a significant advantage for recognition systems running on resource constrained devices and communication channels.

Finger-vein recognition, a new and advanced biometrics recognition method, is attracting the attention of researchers because of its advantages such as high recognition performance and lesser likelihood of theft and inaccuracies occurring on account of skin condition defects. However, as reported by previous researchers, it is possible to attack a finger-vein recognition system by using presentation attack (fake) finger-vein images. As a result, spoof detection, named as presentation attack detection (PAD), is necessary in such recognition systems. Previous attempts to establish PAD methods primarily focused on designing feature extractors by hand (handcrafted feature extractor) based on the observations of the researchers about the difference between real (live) and presentation attack finger-vein images. Therefore, the detection performance was limited. Recently, the deep learning framework has been successfully applied in computer vision and delivered superior results compared to traditional handcrafted methods on various computer vision applications such as image-based face recognition, gender recognition and image classification. In this paper, we propose a PAD method for near-infrared (NIR) camera-based finger-vein recognition system using convolutional neural network (CNN) to enhance the detection ability of previous handcrafted methods. Using the CNN method, we can derive a more suitable feature extractor for PAD than the other handcrafted methods using a training procedure. We further process the extracted image features to enhance the presentation attack finger-vein image detection ability of the CNN method using principal component analysis method (PCA) for dimensionality reduction of feature space and support vector machine (SVM) for classification. Through extensive experimental results, we confirm that our proposed method is adequate for presentation attack finger-vein image detection and it can deliver superior detection results compared to CNN-based methods and other previous handcrafted methods.

A typical automated face recognition system is composed of three main component tasks: face detection and alignment (FDA), face presentation attack detection (FPAD), and face representation and matching (FRM). These tasks are often treated as standalone problems and deep neural network (DNN)-based solutions have been proposed to address them individually. However, in resource-constrained sce-narios it would be ideal to have a unified DNN model that can perform all the three tasks together. As a first step towards realizing this goal, this work attempts to perform joint FRM and FPAD based on a single Vision Transformer (ViT) backbone. Recent work demonstrating the ability of ViT to extract a diverse set of feature representations gives rise to the tantalising possibility of building an end-to-end face recognition system using a single ViT model. The standard approach for designing multi-task DNNs is to implement different classification heads (e.g., for FRM and FPAD) based on a common stem/base and learn these heads either individually or jointly. A key contribution of this work is to demonstrate that this naive multi-head approach results in sub-optimal performance for either FRM or FPAD, because the features required by these tasks are very different. While good FPAD performance depends on accurately characterizing the micro textures, face matching demands attention towards more global characteris-tics. Hence, we propose a novel feature ensemble approach, where an ensemble of local features extracted from the in-termediate blocks of a ViT are utilized for FPAD, while face matching is performed based on the ViT class token. Exper-iments demonstrate that the proposed ViT feature ensemble approach is able to achieve good performance for both face matching and FPAD compared to the multi-head approach.

Individual recognition systems with high-speed and high-accuracy have been recently demanded in the automatic logging into a PC, the immigration at the airport, the access control and diligence & indolence management in an office, and so on. Biometric authentication is now being regarded as the most valid method because of the receptivity, individuality and invariability of biometric identifiers. Various types of the individual recognition systems based on biometrics have been studied and partially realized. Fingerprints, faces, hand geometry, irises, vein patterns, gait, signatures, etc., are known as biometric identifiers. In particular, the fingerprint recognition system has been widely used because of its high reliability and reasonable price (Maltoni et al., 2003a; Jain et al., 2010). The fingerprint recognition methods can be classified into the following three types: (i) the minutiae-based, (ii) the frequency-based and (iii) the image-based methods. The minutiae-based method is mainly being used in the practical fingerprint recognition system. For example, a memetic fingerprint matching algorithm has been recently proposed to identify the optimal or near optimal global matching between two minutiae sets (Sheng et al., 2007). A matching technique for fingerprint recognition using the Minutia CylinderCode (MCC), which is based on 3D data structures built from minutiae distances and angles, has also been made a proposal (Cappelli et al., 2010) to exclude the drawbacks in the fingerprint authentication using local minutiae structures. Moreover, a fingerprint verification using spectral minutiae representations has been suggested to overcome translation, rotation and scaling which are the drawbacks of minutiae-based algorithms (Xu et al., 2009a, 2009b). The frequency-based method, such as the frequency analysis method, is also being used in the practical fingerprint recognition system in order to secretly hide the original fingerprint information (Takeuchi et al., 2007). Recently, a fingerprint recognition method based on melfrequency cepstral coefficients and polynomial shape coefficients has been proposed because of the robustness to noise and the insensivity to translation (Hashad et al., 2010). The image-based method has been being studied to improve the accuracy in the fingerprint recognition. For example, an enhanced image-based algorithm for fingerprint verification based on invariant moment features has been recently proposed to improve matching accuracy and processing speed (Yang & Park, 2008a, 2008b). A novel image-based fingerprint matcher based on the minutiae alignment has also been made a proposal to improve the verification performance (Nanni & Lumini, 2009).

This paper deals with the fingerprint classification and recognition system, which consists of extracting and matching of minutiae from the input image. The necessity for security in fields such as improving airport security, strengthening the national borders, in travel documents, in preventing ID theft has brought the need to develop an able and efficient method for correct classification of personnel authentication. Fingerprints are the first biometric science used widely for the validation and verifications of entry into specific task, which is more reliable, efficient and accurate. Despite finger print recognition being reliable it has disadvantages such as very low recognition rate, low accuracy rate, total time of recovery and data insufficiency. To address above problem a novel data mining technique, neuro nearest neighbour based fingerprint classification and recognition, is introduced which boosts the classification rate. The proposed method consists of different stages such as image enhancement, line detector based feature extraction, neural network classification using Learning vector quantization and Back propagation networks. The proposed system is trained and tested on Fingerprint Database obtained from the University of Bologna Italy, which consists of 900 samples. The exact image is recognized precisely from the classified database rather than the original set of database using Crisp K-nearest neighbour algorithm that increases recognition accuracy and reduction in time

With the rapid increase in fingerprint-based identification systems, there are growing concerns related to spoofs or artificially created fingerprint substantiations that can breach the security of these systems. Several issues associated with the susceptibility of fingerprint authentication systems to spoofing attacks are being underlined in recent research works. To produce spoof artefacts, easily accessible materials like Play-Doh, Gelatin, Wood glue and Ecoflex are etched with ridges of fingerprint modality and presented to the biometric device to circumvent the sensor module. A number of anti-spoofing techniques that incorporate the potent fingerprint features are being developed to countermeasure the fingerprint presentation attacks (PAs). These strategies are intended to thwart an adversary from obtaining unauthorized access to the sensitive information that the fingerprint recognition system is supposed to protect. This paper examines fingerprint presentation attack detection (PAD) mechanisms in depth, with an emphasis on deep learning-based solutions. The review clearly reveals that software-based PAD techniques are more efficient as compared to their hardware counterparts in terms of performance metrics. However, designing fingerprint PAD techniques which provide better performance in cross-datasets and cross-sensor environment still remains an open research problem. Besides, requirement of large data volumes for training and the emergence of novel and easily accessible materials for creating spoof artefacts is still an open research challenge.

A facial recognition system is a technology capable of identifying or verifying a person from a digital image or a video frame from a video source. There are multiple methods in which facial recognition systems work, but in general, they work by comparing selected facial features from the given image with faces within a database. It is also described as a Biometric Artificial Intelligence-based application that can uniquely identify a person by analyzing patterns based on the person's facial textures and shape. While initially a form of computer application, it has seen wider uses in recent times on mobile platforms and in other forms of technology, such as robotics. It is typically used as access control in security systems and can be compared to other biometrics such as fingerprint or eye iris recognition systems. Although the accuracy of facial recognition system as biometric technology is lower than iris recognition and fingerprint recognition, it is widely adopted due to its contactless and non-invasive process. Recently, it has also become popular as a commercial identification and marketing tool. Other applications include advanced human-computer interaction, video surveillance, automatic indexing of images and video database, among others. The use of facial recognition has recently become a very debatable subject and has been criticized more and more because it was considered an unethical tool used to spy on the public. The reason for such criticism is, however, largely due to the lack of information and regulation of this technology. Used proportionally and responsibly, facial recognition can and should be beneficial. It has the capacity to do much more to increase security in the future-from street crime to airport security. Armed war crime has dominated UK titles throughout the year. Recent statistics indicate that the number of people who benefited from the emergency assistance due to armed attacks increased by almost 40% compared to the two years ago, while the number of children under 18 years of age with stab wounds is increasing by 86 % in only four years. Face recognition has become a normal activity in many airports around the world. Many people today have a so-called biometric passport that allows them to go faster to the gate without having to be controlled. The facial recognition used in this way has significantly reduced waiting times for passport control but also has the ability to increase security in and around airports. Face recognition thus allows officers to identify an individual more quickly and accurately than the human eye. While some critics may worry about technology-related confidentiality issues, airports have shown that the use of facial recognition has improved security as well as speeding up processes such as check-in and in the future, even procedures boarding. When used correctly and proportionally, facial recognition can help protect the public and improve national security on multiple fronts. Similarly, advanced technology can recognize a person seen on CCTV security systems at the crime scene, justifying a person's stop and search. The ability to check in real time whether a person is on the list of criminal investigations adds an added advantage to the decision-making process before stopping and searching, thus lowering the probability of discrimination.

There are various types of applications for fingerprint recognition which is used for different purposes. Fingerprint is one of the challenging pattern Recognition problem. The Fingerprint Recognition system is divided into four stages. First is Acquisition stage to capture the fingerprint image, The second is Pre-processing stage to enhancement, binarization, thinning fingerprint image. The Third stage is Feature Extraction Stage to extract the feature from the thinning image by use minutiae extractor methods to extract ridge ending and ridge bifurcation from thinning. The fourth stage is matching(Identification, Verification) to match two minutiae points by using minutiae matcher method in which similarity and distance measure are used. The algorithm is tested accurately and reliably by using fingerprint images from different databases. In this paper the fingerprint databases used are FVC2000 and FVC2002 Databases, we see that, the FVC2002 database perform better results compare with FVC2000 database. The recognition system evaluate with two factor FAR and FRR, In this system the result of FAR is 0.0154 and FRR is 0.0137 with Accuracy equal to 98.55%.

ABSTRACTRecent advances in information and communication technology (ICT) have pushed education systems into the online paradigm. Although fingerprint recognition has received considerable attention from researchers, the concept of creating online or distance education systems for fingerprint recognition systems is a valid problem that requires further investigation. This article addresses the problem and presents a web platform for online education of a fingerprint recognition system. This study focused on the technical design and implementation of the web platform solely as a proof of concept to illustrate the possibility of creating distance education for fingerprint recognition. The developed platform offers students a way to gain knowledge about the full range of fingerprint recognition system technology. The significance of using a web platform stems from the need to supplement students’ theoretical knowledge with practical hands-on experience with information security in general and with fingerprint recognition in particular. Demonstrations such as fingerprint image enhancement and fingerprint feature extraction are included in the current platform prototype. Plots of fingerprint recognition performance curves such as false match rate (FMR) and false non-match rate (FNMR) as functions of the matching threshold curve, the receiver operating characteristic (ROC) curve, and the detection error trade-off (DET) curve are plotted as part of the platform. From a technical viewpoint, the platform was successfully implemented; however, its educational impact needs further evaluation.

This chapter presents a comprehensive study on the application of stimulated pupillary light reflex to presentation attack detection (PAD) that can be used in iris recognition systems. A pupil, when stimulated by visible light in a predefined manner, may offer sophisticated dynamic liveness features that cannot be acquired from dead eyes or other static objects such as printed contact lenses, paper printouts, or prosthetic eyes. Modeling of pupil dynamics requires a few seconds of observation under varying light conditions that can be supplied by a visible light source in addition to the existing near-infrared illuminants used in iris image acquisition. The central element of the presented approach is an accurate modeling and classification of pupil dynamics that makes mimicking an actual eye reaction difficult. This chapter discusses new data-driven models of pupil dynamics based on recurrent neural networks and compares their PAD performance to solutions based on the parametric Clynes–Kohn model and various classification techniques. Experiments with 166 distinct eyes of 84 subjects show that the best data-driven solution, one based on long short-term memory, was able to correctly recognize 99.97% of attack presentations and 98.62% of normal pupil reactions. In the approach using the Clynes–Kohn parametric model of pupil dynamics, we were able to perfectly recognize abnormalities and correctly recognize 99.97% of normal pupil reactions on the same dataset with the same evaluation protocol as the data-driven approach. This means that the data-driven solutions favorably compare to the parametric approaches, which require model identification in exchange for a slightly better performance. We also show that observation times may be as short as 3 s when using the parametric model, and as short as 2 s when applying the recurrent neural network without substantial loss in accuracy. Along with this chapter we also offer: (a) all time series representing pupil dynamics for 166 distinct eyes used in this study, (b) weights of the trained recurrent neural network offering the best performance, (c) source codes of the reference PAD implementation based on Clynes–Kohn parametric model, and (d) all PAD scores that allow the reproduction of the plots presented in this chapter. To our best knowledge, this chapter proposes the first database of pupil measurements dedicated to presentation attack detection and the first evaluation of recurrent neural network-based modeling of pupil dynamics and PAD.

In general, most fingerprint recognition systems are based on the minutiae feature points. When matching two fingerprint images, the goal in most recognition systems is to find the optimal transformation model that aligns their feature points in order to find among them the number of matched or aligned points and then generate a matching score. A major problem in feature extraction stage is that when the fingerprint image is of a poor quality due to skin conditions and sensor noise, that leads to many broken ridges in the image caused by cutline. In this case, the extraction of minutiae leads to a lot of spurious points and the performance of the system will degrade. Usually, image enhancement techniques are applied as preprocessing step to overcome this problem. In this work, we propose to use corner points on fingerprint ridges as new features in addition to the ridges minutiae in order to improve the recognition performance. Every ridge is decomposed into several straight edges (SEs). A straight edge is defined as a straight link of ridge points. On a ridge, the head of the first straight edge and the tail of the last one are two minutia and the intersections of the SEs are the ridge corners. Thus, we propose to use a ridge as primitive rather than individual points for matching. This primitive is a structure consisting of groups of both feature points which are minutiae and corners belonging to the same ridge. Based on this primitive, an intelligent matching technique is introduced using sets of feature points on the same primitive. As a result, the recognition performance is increased since it is based on ridge primitive matching rather than individual minutiae matching. Finally, our experimental results compared with those obtained by other well-known techniques in the literature demonstrate the effectiveness and efficiency of our proposed algorithm.

최근 보안 시스템 중에서 지문인식을 이용한 보안 시스템은 다른 보안 시스템에 비해 유일성과 편의성 등의 장점을 가짐으로써 많은 사람들이 관심을 갖는 분야이다. 지문인식 시스템에 있어서 가장 중요한 사항은 실제 지문과 영상을 통해 얻어진 지문간의 정합에서의 정확성과 지문 인식을 위해 사용하는 영상처리 알고리즘들을 신속하게 처리하는 데 있다. 기존의 지문인식 시스템은 특징 추출을 위해 사용하는 알고리즘들의 처리 시간을 줄이기 위해 전체 처리과정 중 일부 과정들을 생략함으로써 처리시간을 단축한다. 하지만 이 방식은 처리시간을 단축시킬 수 있는 반면 특징 추출에 대한 정확도가 떨어진다. 따라서 본 논문에서는 특징 추출에 대한 정확도를 높이기 위해 전체 처리 과정을 사용하면서 동시에 처리시간도 단축시킬 수 있는 다중 접근 메모리 시스템을 이용한 지문인식 특징 추출 알고리즘을 구현하였고, 구현된 시스템을 사용하였을때 다중접근 메모리 시스템과 시리얼 프로세서의 결과에 대해 correlation을 이용한 검증을 통해 제안된 방법의 신뢰도를 확인하였으며, 시리얼 프로세서에 비해 MAMS-PP64를 이용한 방법은 수행시간에서 약 1.56배 향상되었음을 확인하였다. Among the recent security systems, security system with fingerprint recognition gets many people's interests through the strengths such as exclusiveness, convenience, etc, in comparison with other security systems. The most important matters for fingerprint recognition system are reliability of matching between the fingerprint in database and user's fingerprint and rapid process of image processing algorithms used for fingerprint recognition. The existing fingerprint recognition system reduces the processing time by removing some processes in the feature extraction algorithms but has weakness of a reliability. This paper realizes the fingerprint recognition algorithm using MAMS(Multi-Access Memory System) for both the rapid processing time and the reliability in feature extraction and matching accuracy. Reliability of this process is verified by the correlation between serial processor's results and MAMS-PP64's results. The performance of the method using MAMS-PP64 is 1.56 times faster than compared serial processor.

Automated Fingerprint Recognition Systems (AFRSs) have been threatened by Presentation Attack (PA) since its existence. It is thus desirable to develop effective presentation attack detection (PAD) methods. However, the unpredictable PAs make PAD be a challenging problem. This paper proposes a novel One-Class PAD (OCPAD) method for Optical Coherence Technology (OCT) images based fingerprint PA detection. The proposed OCPAD model is learned from a training set only consists of Bonafides (i.e. real fingerprints). The reconstruction error and latent code obtained from the trained auto-encoder network in the proposed model is taken as the basis for the following spoofness score calculation. To get more accurate reconstruction error, we propose an activation map based weighting model to further refine the accuracy of reconstruction error. We test different statistics and distance measures and finally use a decision level fusion to make the final prediction. Our experiments are performed using a dataset with 93200 bonafide scans and 48400 PA scans. The results show that the proposed OCPAD can achieve a True Positive Rate (TPR) of 99.43% when the False Positive Rate (FPR) equals to 10% and a TPR of 96.59% when FPR=5%, which significantly outperformed a feature based approach and a supervised learning based model requiring PAs for training.

Deep generative variational autoencoding for replay spoof detection in automatic speaker verification

Biometric-based system is one of the systems for identification/recognition and verification. There are many types of biometric technology, one of them is fingerprint recognition system. Fingerprint, like other biometric feature is unique, means that it will always different with its kind. One of the problems in fingerprint recognition systems is how to extract the characteristics/feature of a person's fingerprint. This article present fingerprint matching application as part of recognition, using simple minutiae point extraction algorithm implemented it into hardware-based Artificial Neural Network device. The Cogniblox is an ANN device from CogniMem that comes with 4 CM1K chip, able to store up to 4096 neurons in the device. MATLAB environment is used for developing the fingerprint recognition application. The designed system is then undergo a performance evaluation by recognizing a total of 80 dataset of fingerprint, consist of 10 different fingerprint with 8 impression each, which stored into the neurons beforehand. The performance evaluation resulted in a False Acceptance Rate (FAR), False Rejection Rate (FRR) and Equal Error Rate (EER) are presented.