Dissertations / Theses on the topic 'Liveness detection'

Biometrics deals with identifying individuals with help of their biological data. Fingerprint scanning is the most common method of the biometric methods available today. The security of fingerprint scanners has however been questioned and previous studies have shown that fingerprint scanners can be fooled with artificial fingerprints, i.e. copies of real fingerprints. The fingerprint recognition systems are evolving and this study will discuss the situation of today.

Two approaches have been used to find out how good fingerprint recognition systems are in distinguishing between live fingers and artificial clones. The first approach is a literature study, while the second consists of experiments.

A literature study of liveness detection in fingerprint recognition systems has been performed. A description of different liveness detection methods is presented and discussed. Methods requiring extra hardware use temperature, pulse, blood pressure, electric resistance, etc., and methods using already existent information in the system use skin deformation, pores, perspiration, etc.

The experiments focus on making artificial fingerprints in gelatin from a latent fingerprint. Nine different systems were tested at the CeBIT trade fair in Germany and all were deceived. Three other different systems were put up against more extensive tests with three different subjects. All systems werecircumvented with all subjects'artificial fingerprints, but with varying results. The results are analyzed and discussed, partly with help of the A/R value defined in this report.

This thesis is concerned with liveness detection for biometric systems and in particular for face recognition systems. Biometric systems are well studied and have the potential to provide satisfactory solutions for a variety of applications. However, presentation attacks (spoofng), where an attempt is made at subverting them system by making a deliberate presentation at the sensor is a serious challenge to their use in unattended applications. Liveness detection techniques can help with protecting biometric systems from attacks made through the presentation of artefacts and recordings at the sensor. In this work novel techniques for liveness detection are presented using gaze information. The notion of natural gaze stability is introduced and used to develop a number of novel features that rely on directing the gaze of the user and establishing its behaviour. These features are then used to develop systems for detecting spoofng attempts. The attack scenarios considered in this work include the use of hand held photos and photo masks as well as video reply to subvert the system. The proposed features and systems based on them were evaluated extensively using data captured from genuine and fake attempts. The results of the evaluations indicate that gaze-based features can be used to discriminate between genuine and imposter. Combining features through feature selection and score fusion substantially improved the performance of the proposed features.

The main topic ofmy research during these three years concerned biometrics and in particular the Fingerprint Liveness Detection (FLD), namely the recognition of fake fingerprints. Fingerprints spoofing is a topical issue as evidenced by the release of the latest iPhone and Samsung Galaxy models with an embedded fingerprint reader as an alternative to passwords. Several videos posted on YouTube show how to violate these devices by using fake fingerprints which demonstrated how the problemof vulnerability to spoofing constitutes a threat to the existing fingerprint recognition systems. Despite the fact that many algorithms have been proposed so far, none of them showed the ability to clearly discriminate between real and fake fingertips. In my work, after a study of the state-of-the-art I paid a special attention on the so called textural algorithms. I first used the LBP (Local Binary Pattern) algorithm and then I worked on the introduction of the LPQ (Local Phase Quantization) and the BSIF (Binarized Statistical Image Features) algorithms in the FLD field. In the last two years I worked especially on what we called the “user specific” problem. In the extracted features we noticed the presence of characteristic related not only to the liveness but also to the different users. We have been able to improve the obtained results identifying and removing, at least partially, this user specific characteristic. Since 2009 the Department of Electrical and Electronic Engineering of the University of Cagliari and theDepartment of Electrical and Computer Engineering of the ClarksonUniversity have organized the Fingerprint Liveness Detection Competition (LivDet). I have been involved in the organization of both second and third editions of the Fingerprint Liveness Detection Competition (LivDet 2011 and LivDet 2013) and I am currently involved in the acquisition of live and fake fingerprint that will be inserted in three of the LivDet 2015 datasets.

Face recognition is a mature and reliable technology for identifying people. Due to high-definition cameras and supporting devices, it is considered the fastest and the least intrusive biometric recognition modality. Nevertheless, effective spoofing attempts on face recognition systems were found to be possible. As a result, various anti-spoofing algorithms were developed to counteract these attacks. They are commonly referred in the literature a liveness detection tests. In this research we highlight the effectiveness of some simple, direct spoofing attacks, and test one of the current robust liveness detection algorithms, i.e. the logistic regression based face liveness detection from a single image, proposed by the Tan et al. in 2010, against malicious attacks using processed imposter images. In particular, we study experimentally the effect of common image processing operations such as sharpening and smoothing, as well as corruption with salt and pepper noise, on the face liveness detection algorithm, and we find that it is especially vulnerable against spoofing attempts using processed imposter images. We design and present a new facial database, the Durham Face Database, which is the first, to the best of our knowledge, to have client, imposter as well as processed imposter images. Finally, we evaluate our claim on the effectiveness of proposed imposter image attacks using transfer learning on Convolutional Neural Networks. We verify that such attacks are more difficult to detect even when using high-end, expensive machine learning techniques.

Tato práce se zabývá rozšířením snímače otisků prstů Touchless Biometric Systems 3D-Enroll o jednotku detekce živosti prstu na základě žil. Bylo navrhnuto a zkonstruováno hardwarové řešení s využitím infračervených diod. Navržené softwarové řešení pracuje ve dvou různých režimech: detekce živosti na základě texturních příznaků a verifikace uživatelů na základě porovnávání žilních vzorů. Datový soubor obsahující přes 1100 snímků jak živých prstů tak jejich falsifikátů vznikl jako součást této práce a výkonnost obou zmíněných režimů byla vyhodnocena na tomto datovém souboru. Na závěr byly navrhnuty materiály vhodné k výrobě falsifikátů otisků prstů umožňující oklamání detekce živosti pomocí žilních vzorů.

Thesis (M.S.)--West Virginia University, 2003
Title from document title page. Document formatted into pages; contains viii, 96 p. : ill. (some col.) Includes abstract. Includes bibliographical references (p. 72-74).

Liveness detection in automatic fingerprint identification systems (AFIS) is an issue which still prevents its use in many unsupervised security applications. In the last decade, various hardware and software solutions for the detection of liveness from fingerprints have been proposed by academic research groups. However, the proposed methods have not yet been practically implemented with existing AFIS. A large amount of research is needed before commercial AFIS can be implemented. In this research, novel active pore based liveness detection methods were proposed for AFIS. These novel methods are based on the detection of active pores on fingertip ridges, and the measurement of ionic activity in the sweat fluid that appears at the openings of active pores. The literature is critically reviewed in terms of liveness detection issues. Existing fingerprint technology, and hardware and software solutions proposed for liveness detection are also examined. A comparative study has been completed on the commercially and specifically collected fingerprint databases, and it was concluded that images in these datasets do not contained any visible evidence of liveness. They were used to test various algorithms developed for liveness detection; however, to implement proper liveness detection in fingerprint systems a new database with fine details of fingertips is needed. Therefore a new high resolution Brunel Fingerprint Biometric Database (B-FBDB) was captured and collected for this novel liveness detection research. The first proposed novel liveness detection method is a High Pass Correlation Filtering Algorithm (HCFA). This image processing algorithm has been developed in Matlab and tested on B-FBDB dataset images. The results of the HCFA algorithm have proved the idea behind the research, as they successfully demonstrated the clear possibility of liveness detection by active pore detection from high resolution images. The second novel liveness detection method is based on the experimental evidence. This method explains liveness detection by measuring the ionic activities above the sample of ionic sweat fluid. A Micro Needle Electrode (MNE) based setup was used in this experiment to measure the ionic activities. In results, 5.9 pC to 6.5 pC charges were detected with ten NME positions (50μm to 360 μm) above the surface of ionic sweat fluid. These measurements are also a proof of liveness from active fingertip pores, and this technique can be used in the future to implement liveness detection solutions. The interaction of NME and ionic fluid was modelled in COMSOL multiphysics, and the effect of electric field variations on NME was recorded at 5μm -360μm positions above the ionic fluid.

Abstract. Face anti-spoofing plays a vital role in security systems including face payment systems and face recognition systems. Previous studies showed that live faces and presentation attacks have significant differences in both remote photoplethysmography (rPPG) and texture information. We propose a generalized method exploiting both rPPG and texture features for face anti-spoofing task. First, we design multi-scale long-term statistical spectral (MS-LTSS) features with variant granularities for the representation of rPPG information. Second, a contextual patch-based convolutional neural network (CP-CNN) is used for extracting global-local and multi-level deep texture features simultaneously. Finally, weight summation strategy is employed for decision level fusion of the two types of features, which allow the proposed system to be generalized for detecting not only print attack and replay attack, but also mask attack. Comprehensive experiments were conducted on five databases, namely 3DMAD, HKBU-Mars V1, MSU-MFSD, CASIA-FASD, and OULU-NPU, to show the superior results of the proposed method compared with state-of-the-art methods.Tiivistelmä. Kasvojen anti-spoofingilla on keskeinen rooli turvajärjestelmissä, mukaan lukien kasvojen maksujärjestelmät ja kasvojentunnistusjärjestelmät. Aiemmat tutkimukset osoittivat, että elävillä kasvoilla ja esityshyökkäyksillä on merkittäviä eroja sekä etävalopölymografiassa (rPPG) että tekstuuri-informaatiossa, ehdotamme yleistettyä menetelmää, jossa hyödynnetään sekä rPPG: tä että tekstuuriominaisuuksia kasvojen anti-spoofing -tehtävässä. Ensinnäkin rPPG-informaation esittämiseksi on suunniteltu monivaiheisia pitkän aikavälin tilastollisia spektrisiä (MS-LTSS) ominaisuuksia, joissa on muunneltavissa olevat granulariteetit. Toiseksi, kontekstuaalista patch-pohjaista konvoluutioverkkoa (CP-CNN) käytetään globaalin paikallisen ja monitasoisen syvään tekstuuriominaisuuksiin samanaikaisesti. Lopuksi, painoarvostusstrategiaa käytetään päätöksentekotason fuusioon, joka auttaa yleistämään menetelmää paitsi hyökkäys- ja toistoiskuille, mutta myös peittää hyökkäyksen. Kattavat kokeet suoritettiin viidellä tietokannalla, nimittäin 3DMAD, HKBU-Mars V1, MSU-MFSD, CASIA-FASD ja OULU-NPU, ehdotetun menetelmän parempien tulosten osoittamiseksi verrattuna uusimpiin menetelmiin.

Security breaches due to misidentification of an individual pose one of the greatest threats and challenges for today’s world. The use of biometrics can be quite promising in minimising this threat. Biometrics refers to the automatic authentication of individuals based on their physiological and behavioural characteristics. To date, various biometric systems have been proposed in the literature, among them biometric traits such as the face, iris, fingerprints, retina, gait, and vocal patterns are found to be distinctive to each and every person and are considered to be most reliable biometric identifiers. Regardless of the available biometrics traits, to date, no biometric system has been found to be a perfect, and which can be applied universally in a way that is robust/adaptive to change in different environmental conditions. Multimodal biometric systems were proposed in the late 1990’s to extend the range of biometric applicability. In a multimodal biometric system, two or more biometric identifiers are fused by an information fusion technique, thereby providing robustness for changing in a greater range of environmental conditions and enhancing other properties that an ideal biometric system should possess. Another important property that a biometric system should possess is a capability to distinguish between real and fake data. Although both the robustness of the system and capability to distinguish between a real and fake data should be incorporated into a single system, there is a trade-off. Therefore, due to the aforementioned research problems, this thesis addresses advancements in multimodal ocular biometrics using iris and sclera and also investigates the trade-off between robustness/adaptability and anti-spoofing/liveness detection (which is one method to distinguish between real and fake data). Biometrics traits that allow personal identification, eye traits offer a good choice of biometrics, as the eye offers a wide range of unique characteristics. The two common eye biometric identifiers that can be found in the literature are the iris and retina. Two more biometrics that are becoming popular nowadays are the sclera and the peri-ocular. The iris biometric is believed to be the most reliable eye biometric and that is why various commercial products based on this biometric are available; but the iris biometric used in an unconstrained scenario is still an open research area. The performance of iris biometrics with changes in the gaze angle of the eye can be affected highly. Therefore, due to this restriction, high user cooperation is required by persons with squinty eyes to get successfully identified in an iris biometric system. Identifying individuals with darker irises is another big challenge in iris recognition in the visible spectrum. To mitigate this problem, multi-modal eye biometrics was proposed by combining iris and sclera traits in the visible spectrum. However, in order to establish the concept of multimodal eye biometrics using the iris and sclera, it is first necessary to assess if sufficient discriminatory information can be gained from the sclera, further assessment in regards to its combination with the iris pattern and adaptiveness of the traits with respect to changes in environmental conditions, population, the data acquisition technique and time span. Multimodal biometrics using sclera and iris have not been extensively studied and little is known regarding their usefulness. So, the state-of-the-art related to it is not sufficiently mature and still in its infancy. This thesis concentrates on designing an image processing and pattern recognition module for evaluating the potential of the scleral biometric with regards to biometric accuracy. Thus, research is also carried out investigate usefulness of the sclera trait in combination with the iris pattern. Various, pre-processing techniques, segmentation, feature extraction, information fusion and classification techniques are employed to push the border of this multimodal biometrics. The latter half of the thesis concentrates on bridging the anti-spoofing technique liveliness with adaptiveness of biometrics. Traditional biometric systems are not equipped to distinguish between fake and real data that has been scanned in front of the sensors. As a result, they adhere to forgery attacks by intruders who can take the privilege of a genuine user. With the rising demand of involuntary or unmanned biometric systems in border security, flight checking, and other restricted zones, the incorporation of the automatic detection of forgery attacks is becoming very obvious. Adaptability of the system with respect to the change in the trait is another important aspect that this biometric system should be enriched with. As mentioned previously both the forgery detection method (termed as liveness detection in the literature of biometrics) and adaptability of the trait is necessary for a trusted involuntary biometric system, but initial studies in the literature exhibit it as a trade-off. Therefore to fulfil the gap, this thesis aimed to propose a new framework for software-based liveness detection, which is also associated to the adaptability of the trait. To fulfil the above-highlighted aim in the proposed framework, intra-class level (i.e. user level) liveness detection is introduced employing image quality-based features. Furthermore, to incorporate the adaptability of the trait, online learning-based classifiers are used. Initial investigation and experimental results solicit the use of the proposed framework for trusted involuntary biometric systems. Two new multi-angle eye datasets were developed and published as a part of the current research. The thesis also consists of contributions to other fields of pattern recognition such as wrist vein biometrics, multiscript signature verification and script identification.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Info & Comm Tech
Science, Environment, Engineering and Technology
Full Text

Chapter One contains the introduction, sets the context and aims of the research and highlights the original contribution to knowledge along with publications gained from this research. Chapter Two will cover the methodological choice, which is grounded theory, as well as highlighting other potential methodologies that while are viable have not been used. Chapter Three will identify the background to biometric security while identifying some of the key areas that are currently lacking such as an appropriate way of measuring liveness detection techniques as well as the entire coercion detection sub-discipline. Chapter Four discusses the development of a new taxonomy that will classify liveness detection while moving away from the current ordinal measurement system used within the research area. Analysis of these liveness classifiers will then follow leading into the adaptation of the taxonomy of coercion detection techniques. Further development of these new techniques will follow, identifying metrics for coercion detection and an analysis of the proposed classifiers. After the taxonomy development Chapter Five analyses coercion and liveness techniques by 1.6 Published Work 7 applying the taxonomy across a selection of liveness and coercion techniques. This will then be followed by the development of an algorithm to denote the level of security an individual technique has achieved. Explanation of the algorithm development, components and testing will then be included. Finally Chapter Six will contain the final concluding remarks and will cover some of the areas in the future that can be looked into, alongside some focuses for article and conference submission.

Orientador: Roberto de Alencar Lotufo
Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação
Made available in DSpace on 2018-08-26T03:01:45Z (GMT). No. of bitstreams: 1 Nogueira_RodrigoFrassetto_M.pdf: 3122263 bytes, checksum: e6333eb55b8b4830e318721882159cd1 (MD5) Previous issue date: 2014
Resumo: Com o uso crescente de sistemas de autenticação por biometria nos últimos anos, a detecção de impressões digitais falsas tem se tornado cada vez mais importante. Neste trabalho, nós implementamos e comparamos várias técnicas baseadas em software para detecção de vivacidade de impressões digitais. Utilizamos como extratores de características as redes convolucionais, que foram usadas pela primeira vez nesta área, e Local Binary Patterns (LBP). As técnicas foram usadas em conjunto com redução de dimensionalidade através da Análise de Componentes Principais (PCA) e um classificador Support Vector Machine (SVM). O aumento artificial de dados foi usado de forma bem sucedida para melhorar o desempenho do classificador. Testamos uma variedade de operações de pré-processamento, tais como filtragem em frequência, equalização de contraste e filtragem da região de interesse. Graças aos computadores de alto desempenho disponíveis como serviços em nuvem, foi possível realizar uma busca extensa e automática para encontrar a melhor combinação de operações de pré-processamento, arquiteturas e hiper-parâmetros. Os experimentos foram realizados nos conjuntos de dados usados nas competições Liveness Detection nos anos de 2009, 2011 e 2013, que juntos somam quase 50.000 imagens de impressões digitais falsas e verdadeiras. Nosso melhor método atinge uma taxa média de amostras classificadas corretamente de 95,2%, o que representa uma melhora de 59% na taxa de erro quando comparado com os melhores resultados publicados anteriormente
Abstract: With the growing use of biometric authentication systems in the past years, spoof fingerprint detection has become increasingly important. In this work, we implemented and compared various techniques for software-based fingerprint liveness detection. We use as feature extractors Convolutional Networks with random weights, which are applied for the first time for this task, and Local Binary Patterns. The techniques were used in conjunction with dimensionality reduction through Principal Component Analysis (PCA) and a Support Vector Machine (SVM) classifier. Dataset Augmentation was successfully used to increase classifier¿s performance. We tested a variety of preprocessing operations such as frequency filtering, contrast equalization, and region of interest filtering. An automatic and extensive search for the best combination of preprocessing operations, architectures and hyper-parameters was made, thanks to the fast computers available as cloud services. The experiments were made on the datasets used in The Liveness Detection Competition of years 2009, 2011 and 2013 that comprise almost 50,000 real and fake fingerprints¿ images. Our best method achieves an overall rate of 95.2% of correctly classified samples - an improvement of 59% in test error when compared with the best previously published results
Mestrado
Energia Eletrica
Mestre em Engenharia Elétrica

This work focuses on liveness detection for the fingerprint recognition technology. The first part of this thesis describes biometrics, biometric systems, liveness detection and the method for liveness detection is proposed, which is based on spectroscopic characteristics of human skin. The second part describes and summarizes performed experiments. In the end, the results are discussed and further improvements are proposed.

Abstract Because of its natural and non-intrusive interaction, identity verification and recognition using facial information is among the most active areas in computer vision research. Unfortunately, it has been shown that conventional 2D face recognition techniques are vulnerable to spoofing attacks, where a person tries to masquerade as another one by falsifying biometric data and thereby gaining an illegitimate advantage. This thesis explores different directions for software-based face anti-spoofing. The proposed approaches are divided into two categories: first, low-level feature descriptors are applied for describing the static and dynamic characteristic differences between genuine faces and fake ones in general, and second, complementary attack-specific countermeasures are investigated in order to overcome the limitations of generic spoof detection schemes. The static face representation is based on a set of well-known feature descriptors, including local binary patterns, Gabor wavelet features and histogram of oriented gradients. The key idea is to capture the differences in quality, light reflection and shading by analysing the texture and gradient structure of the input face images. The approach is then extended to the spatiotemporal domain when both facial appearance and dynamics are exploited for spoof detection using local binary patterns from three orthogonal planes. It is reasonable to assume that no generic spoof detection scheme is able to detect all known, let alone unseen, attacks scenarios. In order to find out well-generalizing countermeasures, the problem of anti-spoofing is broken into two attack-specific sub-problems based on whether the spoofing medium can be detected in the provided view or not. The spoofing medium detection is performed by describing the discontinuities in the gradient structures around the detected face. If the display medium is concealed outside the view, a combination of face and background motion correlation measurement and texture analysis is applied. Furthermore, an open-source anti-spoofing fusion framework is introduced and its system-level performance is investigated more closely in order to gain insight on how to combine different anti-spoofing modules. The proposed spoof detection schemes are evaluated on the latest benchmark datasets. The main findings of the experiments are discussed in the thesis
Tiivistelmä Kasvokuvaan perustuvan henkilöllisyyden tunnistamisen etuja ovat luonnollinen vuorovaikutus ja etätunnistus, minkä takia aihe on ollut erittäin aktiivinen tutkimusalue konenäön tutkimuksessa. Valitettavasti tavanomaiset kasvontunnistustekniikat ovat osoittautuneet haavoittuvaisiksi hyökkäyksille, joissa kameralle esitetään jäljennös kohdehenkilön kasvoista positiivisen tunnistuksen toivossa. Tässä väitöskirjassa tutkitaan erilaisia ohjelmistopohjaisia ratkaisuja keinotekoisten kasvojen ilmaisuun petkuttamisen estämiseksi. Työn ensimmäisessä osassa käytetään erilaisia matalan tason piirteitä kuvaamaan aitojen ja keinotekoisten kasvojen luontaisia staattisia ja dynaamisia eroavaisuuksia. Työn toisessa osassa esitetään toisiaan täydentäviä hyökkäystyyppikohtaisia vastakeinoja, jotta yleispätevien menetelmien puutteet voitaisiin ratkaista ongelmaa rajaamalla. Kasvojen staattisten ominaisuuksien esitys perustuu yleisesti tunnettuihin matalan tason piirteisiin, kuten paikallisiin binäärikuvioihin, Gabor-tekstuureihin ja suunnattujen gradienttien histogrammeihin. Pääajatuksena on kuvata aitojen ja keinotekoisten kasvojen laadun, heijastumisen ja varjostumisen eroavaisuuksia tekstuuria ja gradienttirakenteita analysoimalla. Lähestymistapaa laajennetaan myös tila-aika-avaruuteen, jolloin hyödynnetään samanaikaisesti sekä kasvojen ulkonäköä ja dynamiikkaa irroittamalla paikallisia binäärikuvioita tila-aika-avaruuden kolmelta ortogonaaliselta tasolta. Voidaan olettaa, ettei ole olemassa yksittäistä yleispätevää vastakeinoa, joka kykenee ilmaisemaan jokaisen tunnetun hyökkäystyypin, saati tuntemattoman. Näin ollen työssä keskitytään tarkemmin kahteen hyökkäystilanteeseen. Ensimmäisessä tapauksessa huijausapuvälineen reunoja ilmaistaan analysoimalla gradienttirakenteiden epäjatkuvuuksia havaittujen kasvojen ympäristössä. Jos apuvälineen reunat on piilotettu kameran näkymän ulkopuolelle, petkuttamisen ilmaisu toteutetaan yhdistämällä kasvojen ja taustan liikkeen korrelaation mittausta ja kasvojen tekstuurianalyysiä. Lisäksi työssä esitellään vastakeinojen yhdistämiseen avoimen lähdekoodin ohjelmisto, jonka avulla tutkitaan lähemmin menetelmien fuusion vaikutuksia. Tutkimuksessa esitetyt menetelmät on kokeellisesti vahvistettu alan viimeisimmillä julkisesti saatavilla olevilla tietokannoilla. Tässä väitöskirjassa käydään läpi kokeiden päähavainnot

This project deals with general biometrics issues focusing on fingerprint biometrics, with description of dermal papillae and principles of fingerprint sensors. Next this work deals with fingerprint liveness detection issues, including description of methods of detection. Next this work describes chosen features for own detection, used database of fingerprints and own algorithm for image pre-processing. Furthermore neural network classifier for liveness detection with chosen features is decribed followed by statistic evaluation of the chosen features and detection results as well as description of the created graphical user interface.

This work outlines a method of detection of the papillary line width in fingerprints. This method is one of the possible methods of liveness detection. The first part of the work with deals defining of the fingerprint, attacks on today's systems and possibilities to improve security. The next section detection describes of the papillary line width. During the process of resolving, the first thing to do was to start operation of the scanning device and to read the database for tests and experiments. An independent application was created on this purpose. Further, there were projected methods for detection and measuring of the papillary line width. Use of the Canny edge detector with the Sobel operator and the Gaussian filter proved the best. Then, there is described implementation of individual methods. The next part of the work describes and assesses the results of the tests. The last chapter summarizes the work and proposes further possibilities of development.

This master‘s thesis deals with biometric fingerprint liveness detection. The theoretical part of the work describes fingerprint recognition biometric systems, fingerprint liveness detection issues and methods for fingerprint liveness detection. The practical part of the work describes proposed set of discriminant features and preprocessing of fingerprint image. Proposed approach using neural network to detect a liveness. The algorithm is tested on LivDet database comprising real and fake images acquired with tree sensors. Classification performance approximately 93% was obtained.

Hlavním přínosem této práce jsou dva nové přístupy pro zvýšení bezpečnosti biometrických systémů založených na rozpoznávání podle otisků prstů. První přístup je z oblasti testování živosti a znemožňuje použití různých typů falešných otisků prstů a jiných metod oklamání senzoru v průběhu procesu snímání otisků. Tento patentovaný přístup je založen na změně barvy a šířky papilárních linií vlivem přitlačení prstu na skleněný podklad. Výsledná jednotka pro testování živosti může být integrována do optických senzorů.  Druhý přístup je z oblasti standardizace a zvyšuje bezpečnost a interoperabilitu procesů extrakce markantů a porovnání. Pro tyto účely jsem vytvořila metodologii, která stanovuje míry sémantické shody pro extraktory markantů otisků prstů. Markanty nalezené testovanými extraktory jsou porovnávány oproti Ground-Truth markantům získaným pomocí shlukování dat poskytnutých daktyloskopickými experty. Tato navrhovaná metodologie je zahrnuta v navrhovaném dodatku k normě ISO/IEC 29109-2 (Amd. 2 WD4).

There are several frauds against biometric systems (BSs) and several techniques exist to secure BSs against these frauds. One of the techniques is liveness detection. To fool fingerprint sensors, latent fingerprints, dummy fingers and wafer-thin layer attached to the finger are being used. Liveness detection is being used also when scanning fingerprints. Several different characteristics of the live finger can be used to detect liveness, for example sweat, conductivity etc. In this thesis, new approach is examined. It is based on the expandability of the finger as an effect of heartbeats/pulsation. As the skin is expanding, also the distances between papillary lines are expanding. Whole finger expands approximately in range of 4,5 ľm , the distance between two neighbor papillary lines in 0,454 ľm . This value collides with wavelength of blue and green light. The result from this work is following. The resolution of the capturing device is not high enough to capture the expandability on distance between two neighbor papillary lines. Also, because of collision with wavelength, the diffraction effect is presented and the result images are influenced by this error.

Dissertação de Mestrado Integrado em Engenharia Electrotécnica e de Computadores apresentada à Faculdade de Ciências e Tecnologia
O interesse pelo tema detecção de vivacidade tem vindo a aumentar nos últimos anos devido ao desenvolvimento de novas ferramentas e conhecimentos na área de biometria. A detecção de vivacidade é essencial em sistemas de autenticação de identidade para impedir que alguém tenham acesso a informações confidenciais ilegalmente. Os sistemas de autenticação facial existentes podem ser enganados através de uma simples fotografia de um usuário legítimo, que pode ser facilmente obtida por meio de redes sociais, vídeo ou máscaras 3D. Vários algoritmos com o objetivo de deteção de vivacidade têm sido propostos para lidar com este problema, usando diferentes abordagens e numerosas bases de dados.A precisão do reconhecimento de rosto é significativamente melhorada usando redes de neuronais profundas devido à sua capacidade de extrair características profundas de rostos humanos. A combinação de recursos de vivacidade das imagens fornece uma generalização melhor para um classificador de liveness detection facial, aproveitando a fusão de recursos ou abordagens de fusão de modalidades, usada por todos os métodos anti-spoofing facial de última geração.Esta dissertação teve como objetivo a criação de um sistema de Detecção de Vivacidade e Reconhecimento Facial baseado em redes neurais convolucionais (RNC) e utilizando informações obtidas de 3 modalidades diferentes (RGB, Infra-vermelhos e Imagens de profundidade). O uso de múltiplas modalidades neste contexto ainda está a ser explorado, e ainda há muito desenvolvimento necessário para alcançar um sistema de autenticação biométrica perfeito.Para avaliar a precisão do nosso sistema, testamos a nossa rede na base de dados CASIA-SURF, que é uma base de dados com imagens faciais provenientes de diferentes modalidades bastante conhecido, e também criamos uma base de dados especificamente para esta tarefa. Testamos a nossa rede com diferentes arquiteturas e vários módulos e implementamos um sistema biométrico que pode funcionar no mundo real em tempo real. Os resultados foram promissores, mostrando a possibilidade de utilização desse sistema para autenticação de usuários no mundo real.
The interest in the topic of Liveness Detection has been increasing in the past few years due to the development of new tools and knowledge in the area of biometry. Liveness Detection is essential in user authentication systems to stop intruders from gaining access to confidential information illegally. The face authentication systems we have today can be victims to a simple photograph of a legitimate user, which can be easily obtained through social media networks, video replay, or 3D masks.Various face anti-spoofing algorithms have been proposed to tackle this problem, using different approaches, and numerous public face anti-spoofing databases and competitions.Face recognition accuracy is significantly improved using deep learning networks due to their ability to extract human faces’ deep features.The combination of liveness features from the image visual cues provides a better generalization for a face anti-spoofing classifier, taking advantage of the feature fusion or score fusion approach, used by all the state-of-the-art face anti-spoofing measures. This dissertation aimed to create a Liveness Detection and Facial Recognition system based on convolutional neural networks (CNN) and using information collected from 3 different modalities (RGB, Infra-Red, and Depth images). The use of multiple modalities in this context is still being explored, and there is still much development needed to achieve the perfect biometric authentication system.To evaluate the accuracy of our system, we tested with the network on the CASIA-SURF Dataset, which is a well-known multi-modality dataset, and we also created a dataset precisely for this task. We tested our network with different network architectures and modules and implemented a biometric system that can work in the real world in real-time. The results were promising, showing the possibility of using this system for user authentication in the real world

碩士
國立臺灣科技大學
資訊工程系
107
With the development of science and technology, face recognition is now an important technology for authentication in various access control applications, especially used in mobile devices. Unlocking by face has gradually replaced fingerprint identification in some scenarios, which becomes one of the major biometric authentication technology of mobile phones. In a common camera, due to the lack of depth information, it is easy to make fake face images to crack the identification system (e.g., paper printing and screen display) compared with other biological features such as fingerprints and palm prints. Therefore, face liveness detection against spoofing attack using 2D image distortion analysis will be a very important issue in the field of information security. By virtue of the different features between real faces and fake faces, this thesis adopts local binary pattern and 2D image distortion analysis to extract texture information of images, which are used for developing our face liveness detection system against spoofing attack to distinguish fake faces from real faces by a deep neural network. The system employs only a single image captured from a common camera to discriminant real faces and fake faces. In the experiments, three kinds of face spoofing databases are used as subjects of cross-validation. The methods and dataset made by ourselves presented in this thesis can effectively classify the authenticity of human faces. The accuracy of the inside test reaches 99.55%, while that of the outside test attains 95.13%. The experimental results show that our face liveness detection system has high accuracy and generality.