O objetivo deste trabalho é desenvolver um algoritmo capaz de extrair medidas dos dedos e da palma das mãos. Essas imagens serão adquiridas através de um banco de imagens digitalizadas em um scanner de mesa, facilmente encontrado, e processadas para se obter uma curva do contorno da mão, permitindo assim que se extraia informações biométricas das mesmas. As imagens armazenadas no banco serão então associadas a um determinado usuário, visando em um estágio futuro, aplicações na identificação pessoal. / The object of this work is to develop an algorithm to extract fingers and hands palm measures. Those images will be acquired from a database of images digitalized by a scanner, easily founded, and processed to get the hand contour, allowing biometrics informations to be extracted. The images stored in the database will be associated with an user, aiming personal identification in a future stage.

O objetivo deste trabalho é desenvolver um algoritmo capaz de extrair medidas dos dedos e da palma das mãos. Essas imagens serão adquiridas através de um banco de imagens digitalizadas em um scanner de mesa, facilmente encontrado, e processadas para se obter uma curva do contorno da mão, permitindo assim que se extraia informações biométricas das mesmas. As imagens armazenadas no banco serão então associadas a um determinado usuário, visando em um estágio futuro, aplicações na identificação pessoal. / The object of this work is to develop an algorithm to extract fingers and hands palm measures. Those images will be acquired from a database of images digitalized by a scanner, easily founded, and processed to get the hand contour, allowing biometrics informations to be extracted. The images stored in the database will be associated with an user, aiming personal identification in a future stage.

Palmar flexion crease matching is a method for verifying or establishing identity. New methods of palmprint identification, that complement existing identification strategies, or reduce analysis and comparison times, will benefit palmprint identification communities worldwide. To this end, this thesis describes new methods of manual and automated palmar flexion crease identification, that can be used to identify palmar flexion creases in online palmprint images. In the first instance, a manual palmar flexion crease identification and matching method is described, which was used to compare palmar flexion creases from 100 palms, each modified 10 times to mimic some of the types of alterations that can be found in crime scene palmar marks. From these comparisons, using manual palmar flexion crease identification, results showed that when labelled within 10 pixels, or 3.5 mm, of the palmar flexion crease, a palmprint image can be identified with a 99.2% genuine acceptance rate and a 0% false acceptance rate. Furthermore, in the second instance, a new method of automated palmar flexion crease recognition, that can be used to identify palmar flexion creases in online palmprint images, is described. A modified internal image seams algorithm was used to extract the flexion creases, and a matching algorithm, based on kd-tree nearest neighbour searching, was used to calculate the similarity between them. Results showed that in 1000 palmprint images from 100 palms, when compared to manually identified palmar flexion creases, a 100% genuine acceptance rate was achieved with a 0.0045% false acceptance rate. Finally, to determine if automated palmar flexion crease recognition can be used as an effective method of palmprint identification, palmar flexion creases from two online palmprint image data sets, containing images from 100 palms and 386 palms respectively, were automatically extracted and compared. In the first data set, that is, for images from 100 palms, an equal error rate of 0.3% was achieved. In the second data set, that is, for images from 386 palms, an equal error rate of 0.415% was achieved.

The development of accurate and reliable security systems is a matter of wide interest, and in this context biometrics is seen as a highly effective automatic mechanism for personal identification. Among biometric technologies, IrisCode developed by Daugman in 1993 is regarded as a highly accurate approach, being able to support real-time personal identification of large databases. Since 1993, on the top of IrisCode, different coding methods have been proposed for iris and fingerprint identification. In this research, I extend and generalize IrisCode for real-time secure palmprint identification. PalmCode, the first coding method for palmprint identification developed by me in 2002, directly applied IrisCode to extract phase information of palmprints as features. However, I observe that the PalmCodes from the different palms are similar, having many 45o streaks. Such structural similarities in the PalmCodes of different palms would reduce the individuality of PalmCodes and the performance of palmprint identification systems. To reduce the correlation between PalmCodes, in this thesis, I employ multiple elliptical Gabor filters with different orientations to compute different PalmCodes and merge them to produce a single feature, called Fusion Code. Experimental results demonstrate that Fusion Code performs better than PalmCode. Based on the results of Fusion Code, I further identify that the orientation fields of palmprints are powerful features. Consequently, Competitive Code, which uses real parts of six Gabor filters to estimate the orientation fields, is developed. To embed the properties of IrisCode, such as high speed matching, in Competitive Code, a novel coding scheme and a bitwise angular distance are proposed. Experimental results demonstrate that Competitive Code is much more effective than other palmprint algorithms. Although many coding methods have been developed based on IrisCode for iris and palmprint identification, we lack a detailed analysis of IrisCode. One of the aims of this research is to provide such analysis as a way of better understanding IrisCode, extending the coarse phase representation to a precise phase representation, and uncovering the relationship between IrisCode and other coding methods. This analysis demonstrates that IrisCode is a clustering process with four prototypes; the locus of a Gabor function is a two-dimensional ellipse with respect to a phase parameter and the bitwise hamming distance can be regarded as a bitwise angular distance. In this analysis, I also point out that the theoretical evidence of the imposter binomial distribution of IrisCode is incomplete. I use this analysis to develop a precise phase representation which can enhance iris recognition accuracy and to relate IrisCode and other coding methods. By making use of this analysis, principal component analysis and simulated annealing, near optimal filters for palmprint identification are sought. The near optimal filters perform better than Competitive Code in term of d’ index. Identical twins having the closest genetics-based relationship are expected to have maximum similarity in their biometrics. Classifying identical twins is a challenging problem for some automatic biometric systems. Palmprint has been studied for personal identification for many years. However, genetically identical palmprints have not been studied. I systemically examine Competitive Code on genetically identical palmprints for automatic personal identification and to uncover the genetically related palmprint features. The experimental results show that the three principal lines and some portions of weak lines are genetically related features but our palms still contain rich genetically unrelated features for classifying identical twins. As biometric systems are vulnerable to replay, database and brute-force attacks, such potential attacks must be analyzed before they are massively deployed in security systems. I propose projected multinomial distribution for studying the probability of successfully using brute-force attacks to break into a palmprint system based on Competitive Code. The proposed model indicates that it is computationally infeasible to break into the palmprint system using brute-force attacks. In addition to brute-force attacks, I address the other three security issues: template re-issuances, also called cancellable biometrics, replay attacks, and database attacks. A random orientation filter bank (ROFB) is used to generate cancellable Competitive Codes for templates re-issuances. Secret messages are hidden in templates to prevent replay and database attacks. This technique can be regarded as template watermarking. A series of analyses is provided to evaluate the security levels of the measures.

The development of accurate and reliable security systems is a matter of wide interest, and in this context biometrics is seen as a highly effective automatic mechanism for personal identification. Among biometric technologies, IrisCode developed by Daugman in 1993 is regarded as a highly accurate approach, being able to support real-time personal identification of large databases. Since 1993, on the top of IrisCode, different coding methods have been proposed for iris and fingerprint identification. In this research, I extend and generalize IrisCode for real-time secure palmprint identification. PalmCode, the first coding method for palmprint identification developed by me in 2002, directly applied IrisCode to extract phase information of palmprints as features. However, I observe that the PalmCodes from the different palms are similar, having many 45o streaks. Such structural similarities in the PalmCodes of different palms would reduce the individuality of PalmCodes and the performance of palmprint identification systems. To reduce the correlation between PalmCodes, in this thesis, I employ multiple elliptical Gabor filters with different orientations to compute different PalmCodes and merge them to produce a single feature, called Fusion Code. Experimental results demonstrate that Fusion Code performs better than PalmCode. Based on the results of Fusion Code, I further identify that the orientation fields of palmprints are powerful features. Consequently, Competitive Code, which uses real parts of six Gabor filters to estimate the orientation fields, is developed. To embed the properties of IrisCode, such as high speed matching, in Competitive Code, a novel coding scheme and a bitwise angular distance are proposed. Experimental results demonstrate that Competitive Code is much more effective than other palmprint algorithms. Although many coding methods have been developed based on IrisCode for iris and palmprint identification, we lack a detailed analysis of IrisCode. One of the aims of this research is to provide such analysis as a way of better understanding IrisCode, extending the coarse phase representation to a precise phase representation, and uncovering the relationship between IrisCode and other coding methods. This analysis demonstrates that IrisCode is a clustering process with four prototypes; the locus of a Gabor function is a two-dimensional ellipse with respect to a phase parameter and the bitwise hamming distance can be regarded as a bitwise angular distance. In this analysis, I also point out that the theoretical evidence of the imposter binomial distribution of IrisCode is incomplete. I use this analysis to develop a precise phase representation which can enhance iris recognition accuracy and to relate IrisCode and other coding methods. By making use of this analysis, principal component analysis and simulated annealing, near optimal filters for palmprint identification are sought. The near optimal filters perform better than Competitive Code in term of d’ index. Identical twins having the closest genetics-based relationship are expected to have maximum similarity in their biometrics. Classifying identical twins is a challenging problem for some automatic biometric systems. Palmprint has been studied for personal identification for many years. However, genetically identical palmprints have not been studied. I systemically examine Competitive Code on genetically identical palmprints for automatic personal identification and to uncover the genetically related palmprint features. The experimental results show that the three principal lines and some portions of weak lines are genetically related features but our palms still contain rich genetically unrelated features for classifying identical twins. As biometric systems are vulnerable to replay, database and brute-force attacks, such potential attacks must be analyzed before they are massively deployed in security systems. I propose projected multinomial distribution for studying the probability of successfully using brute-force attacks to break into a palmprint system based on Competitive Code. The proposed model indicates that it is computationally infeasible to break into the palmprint system using brute-force attacks. In addition to brute-force attacks, I address the other three security issues: template re-issuances, also called cancellable biometrics, replay attacks, and database attacks. A random orientation filter bank (ROFB) is used to generate cancellable Competitive Codes for templates re-issuances. Secret messages are hidden in templates to prevent replay and database attacks. This technique can be regarded as template watermarking. A series of analyses is provided to evaluate the security levels of the measures.

In this thesis work, a detailed analysis of biometric technologies has been done and a new palmprint recognition algorithm has been implemented. The proposed algorithm is based on 2-D Gabor filters. The developed algorithm is first tested on The Hong Kong Polytechnic University Palmprint Database in terms of accuracy, speed and template size. Then a scanner is integrated into the developed algorithm in order to acquire palm images / in this way an online palmprint recognition system has been developed. Then a small palmprint database is formed via this system in Middle East Technical University. Results on this new database have also shown the success of the developed algorithm.

Current Automatic Palmprint Identification Systems (APIS) closely follow the matching philosophy of Automatic Fingerprint Identification Systems (AFIS), in that they exclusively use a small subset of Level II palmar detail, when matching a latent to an exemplar palm print. However, due the increased size and the significantly more complex structure of the palm, it has long been recognised that there is much detail that remains underutilised. Forensic examiners routinely use this additional information when manually matching latents. The thesis develops novel automatic feature extraction and matching methods which exploit the underutilised Level I detail contained in the friction ridge flow. When applied to a data base of exemplars, the approach creates a ranked list of matches. It is shown that the matching success rate varied with latent size. For latents of diameter 38mm, 91:1% were ranked first and 95:6% of the matches were contained within the ranked top 10. The thesis presents improved orientation field extraction methods which are optimised for friction ridge flow and novel enhancement techniques, based upon the novel use of local circular statistics on palmar orientation fields. In combination, these techniques are shown to provide a more accurate orientation estimate than previous work. The novel feature extraction stages exploit the level sets of higher order local circular statistics, which naturally segment the palm into homogeneous regions representing Level I detail. These homogeneous regions, characterised by their spatial and circular features, are used to form a novel compact tree-like hierarchical representation of the Level I detail. Matching between the latent and an exemplar is performed between their respective tree-like hierarchical structures. The methods developed within the thesis are complementary to current APIS techniques.

La biométrie est une alternative qui se base sur l'identification des personnes à partir de leurs caractéristiques physiques (empreinte digitale, forme de la main, empreinte palmaire) et/ou comportementales (voix, signature dynamique). La biométrie tend à réaliser deux buts importants dans notre vie courante. Le premier but est de réaliser la sécurité en éliminant le doute sur l'identité d'une personne et le second but est de faciliter l'identification des individus. En effet, cette méthode d'identification est de plus en plus préférée par rapport aux méthodes traditionnelles impliquant les mots de passe et les badges. Les travaux de recherche de cette thèse s'inscrivent dans le cadre de la reconnaissance de personnes à l'aide de la biométrie de la main. L'objectif principal est de concevoir un système biométrique multimodal basé sur la fusion de la forme de la main et de l'empreinte palmaire.La première partie de cette thèse propose un nouveau système uni-modal de vérification de la forme de la main. En effet, ce système est basé d'une part, sur la détection du meilleur ensemble des points-clés localisés sur le contour de la main pour adopter la description SIFT (Scale Invariant Feature Transform). D'autre part, un raffinement de correspondance, basé région et apparence de la main est proposé, afin de raffiner autant que possible les points-clés faussement matchés.Tandis que la deuxième partie consiste à proposer un nouveau système d'identification palmaire. En effet, la méthode de représentation parcimonieuse est adoptée afin de décrire le trait biométrique de l'empreinte palmaire. Elle est basée sur l'extraction de descripteurs SIFT de chacun des points-clés détectés. Notre troisième partie concerne la proposition de différentes méthodes de fusion multi-types de la multi modalité, comprenant la fusion multi-représentation, la fusion multi-biométrique et la fusion multi-instance. En effet, la fusion multi-représentation est basée sur la combinaison de descripteurs SIFT et les caractéristiques géométriques de la main au niveau des scores, pour la vérification de la forme de la main. La fusion multi-biométrique est basée sur la combinaison des deux modalités biométriques à savoir la forme de la main et l'empreinte palmaire, au niveau des caractéristiques et de la décision. Par contre, la fusion multi-instance est basée sur la combinaison des empreintes palmaires droite et gauche, au niveau du rang.Ces différentes méthodes de fusion ont prouvé leur efficacité en obtenant de meilleurs taux de reconnaissance, qui sont compétitifs par rapport à d'autres approches multimodales de la biométrie de la main. / Biometry is a technology which is based on the personal identification using their physical features (fingerprint, hand geometry, palmprint) and/or behavioral features (voice, dynamic signature). Biometry aims to achieve two important goals in our current life. The first one is to ensure security by eliminating doubt regarding the identity of a person and the second one is to facilitate the identification of individuals. Indeed, this method of identification is increasingly preferred over traditional methods including passwords and badges. The research works of this thesis talk about the personal recognition using hand biometrics. The main objective is to design a multimodal biometric system based on the fusion of hand shape and palmprint modalities.Our first part is to propose a new unimodal biometric system for hand shape verification. In fact, this system is based firstly, on the detection of the best set of keypoints located on the contour of the hand for further SIFT (Scale Invariant Feature Transform) description. On the other hand, a matching refinement based hand region and appearance is proposed in order to refine as much as possible false matched keypoints.Our second part consists in the proposition of a new palmprint identification system. In fact, the sparse representation method is adopted in order to describe the palmprint biometric trait. It is based on the extraction on SIFT descriptors for each detected keypoint.Our third part concerns the proposition of multi-type fusion methods for multimodality, including the multi-representation fusion, the multi-biometric fusion and the multi-instance fusion. Indeed, the multi-representation fusion method is based on the combination of SIFT descriptors and geometrical features of the hand, at score level. The multi-biometric fusion method is based on the fusion of hand shape and palmprint modalities, at feature and decision levels. On the other hand, the multi-instance fusion method is based on the combination of left and right palmprints, at rank level.These different methods of fusion have proven their effectiveness by achieving encouraging recognition rates that are competitive to other popular multimodal hand biometric approaches.

Submitted by Viviane Lima da Cunha (viviane@biblioteca.ufpb.br) on 2016-02-16T16:06:07Z No. of bitstreams: 1 arquivototal.pdf: 3758225 bytes, checksum: 9b36904ee6da8a9b5182f3af462d91cd (MD5) / Made available in DSpace on 2016-02-16T16:06:07Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 3758225 bytes, checksum: 9b36904ee6da8a9b5182f3af462d91cd (MD5) Previous issue date: 2015-10-23 / Biometrics provides a trusted authentication mechanism by using traits (physical or behavioral) which identify users based on their natural characteristics. Biometric services of classification and verification of users are considered, in principle, more secure than password-based systems, requiring the presentation of a unique physical characteristic and, therefore, the presence of the user at least in the moment of authentication. However, such methods have vulnerabilities that result in high rates of false verification, even in the most modern systems. On the other hand, Genetic algorithms (GA) are an optimization approach based on the principle of natural selection proposed by Charles Darwin which has been proving to be a useful tool in finding solutions to complex problems. Moreover, the use of genetic algorithms in biometrics systems has also been growing as they are an interesting alternative for selecting features. This work applies a genetic algorithm-based approach to optimizing parameters of classification and verification of a hand dataset. The BioPass-UFPB multi-biometric dataset is presented and used to test and validate the proposed method. In total, 99 features – 85 geometric features and 14 texture features - extracted from each hand image were used. Additionally, the importance of each feature is also analyzed. The results showed relative improvements of EER greater than 30% and 90% in the best cases of the two verification approaches performed, respectively. As for classification, the use of genetic algorithms were able to reduce, on average, the number of templates to be recovered by the system to ensure that at least one of these is of the same class of the reference sample. In conclusion, both the results showed and the BioPass-UFPB dataset might help the development of new hand geometry-based biometric recognition systems. / A Biometria oferece um mecanismo de autenticação confiável utilizando traços (físicos ou comportamentais) que permitem identificar usuários baseados em suas características naturais. Serviços biométricos de classificação e verificação de usuários são considerados, a princípio, mais seguros que sistemas baseados em políticas de senha, por exigirem a apresentação de uma característica física única e, portanto, a presença do usuário ao menos no momento da autenticação. No entanto, tais métodos apresentam vulnerabilidades e podem resultar em alta taxa de verificação falsa, mesmo em sistemas mais modernos. Algoritmos genéticos (GA), por sua vez, são uma abordagem de otimização baseada no princípio da seleção natural de Charles Darwin e que vêm provando, ao longo dos anos, ser uma ferramenta útil na busca de soluções em problemas complexos. Além disso, seu uso em sistemas biométricos também vem crescendo por se mostrar uma alternativa para seleção de características. Este trabalho utiliza os algoritmos genéticos como ferramenta na otimização de atributos para classificação e verificação biométrica por geometria da mão e impressão palmar. A Base Multibiométrica BioPass-UFPB é apresentada e empregada para teste e validação do método proposto. Ao todo, 99 atributos – sendo 85 geométricos e 14 de textura - extraídos de cada imagem são utilizados e análises sobre a importância desses atributos são realizadas. Os resultados mostraram que, nas duas abordagens de verificação empregadas, os algoritmos genéticos conseguiram melhoras superiores a 30% e 90% da EER em relação ao caso em que o GA não era aplicado. Na classificação, o uso de algoritmos genéticos conseguiu reduzir na média o número de templates a serem recuperados pelo sistema para garantir que ao menos um desses seja da mesma classe da amostra de referência. Por fim, espera-se que os resultados deste trabalho, bem como a base BioPass-UFPB, sirvam de referência na implementação de novos sistemas de reconhecimento biométrico baseados na geometria da mão.

Combiner les sources d'information pour créer un système de reconnaissance biométrique multimodal permet d'atténuer les limitations de chaque caractéristique utilisée, et donne l'opportunité d'améliorer significativement les performances. Le travail présenté dans ce manuscrit a été réalisé dans le but de proposer un système de reconnaissance performant, qui réponde à des contraintes d'utilisation grand-public, et qui puisse être implanté sur un système matériel de faible coût. La solution choisie explore les possibilités apportées par la multimodalité, et en particulier par la fusion du visage et de la paume. La chaîne algorithmique propose un traitement basé sur les filtres de Gabor, ainsi qu’une fusion des scores. Une base multimodale réelle de 130 sujets acquise sans contact a été conçue et réalisée pour tester les algorithmes. De très bonnes performances ont été obtenues, et ont été confirmées sur une base virtuelle constituée de deux bases publiques (les bases AR et PolyU). L'étude approfondie de l'architecture des DSP, et les différentes implémentations qui ont été réalisées sur un composant de type TMS320c64x, démontrent qu'il est possible d'implanter le système sur un unique DSP avec des temps de traitement très courts. De plus, un travail de développement conjoint d'algorithmes et d'architectures pour l'implantation FPGA a démontré qu'il était possible de réduire significativement ces temps de traitement. / Including multiple sources of information in personal identity recognition reduces the limitations of each used characteristic and gives the opportunity to greatly improve performance. This thesis presents the design work done in order to build an efficient generalpublic recognition system, which can be implemented on a low-cost hardware platform. The chosen solution explores the possibilities offered by multimodality and in particular by the fusion of face and palmprint. The algorithmic chain consists in a processing based on Gabor filters and score fusion. A real database of 130 subjects has been designed and built for the study. High performance has been obtained and confirmed on a virtual database, which consists of two common public biometric databases (AR and PolyU). Thanks to a comprehensive study on the architecture of the DSP components and some implementations carried out on a DSP belonging to the TMS320c64x family, it has been proved that it is possible to implement the system on a single DSP with short processing times. Moreover, an algorithms and architectures development work for FPGA implementation has demonstrated that these times can be significantly reduced.