Anonymization is a broad term. Meaning that personal data, or rather data that identifies a person, is redacted or obscured. In the context of video and image data, the most palpable information is the face. Faces barely change compared to other aspect of a person, such as cloths, and we as people already have a strong sense of recognizing faces. Computers are also adroit at recognizing faces, with facial recognition models being exceptionally powerful at identifying and comparing faces. Therefore it is generally considered important to obscure the faces in video and image when aiming for keeping it anonymized. Traditionally this is simply done through blurring or masking. But this de- stroys useful information such as eye gaze, pose, expression and the fact that it is a face. This is an especial issue, as today our society is data-driven in many aspects. One obvious such aspect is autonomous driving and driver monitoring, where necessary algorithms such as object-detectors rely on deep learning to function. Due to the data hunger of deep learning in conjunction with society’s call for privacy and integrity through regulations such as the General Data Protection Regularization (GDPR), anonymization that preserve useful information becomes important. This Thesis investigates the potential and possible limitation of anonymizing faces without destroying the aforementioned useful information. The base approach to achieve this is through face swapping and face manipulation, where the current research focus on changing the face (or identity) while keeping the original attribute information. All while being incorporated and consistent in an image and/or video. Specifically, will this Thesis demonstrate how target-oriented and subject-agnostic face swapping methodologies can be utilized for realistic anonymization that preserves attributes. Thru this, this Thesis points out several approaches that is: 1) controllable, meaning the proposed models do not naively changes the identity. Meaning that what kind of change of identity and magnitude is adjustable, thus also tunable to guarantee anonymization. 2) subject-agnostic, meaning that the models can handle any identity. 3) fast, meaning that the models is able to run efficiently. Thus having the potential of running in real-time. The end product consist of an anonymizer that achieved state-of-the-art performance on identity transfer, pose retention and expression retention while providing a realism. Apart of identity manipulation, the Thesis demonstrate potential security issues. Specifically reconstruction attacks, where a bad-actor model learns convolutional traces/patterns in the anonymized images in such a way that it is able to completely reconstruct the original identity. The bad-actor networks is able to do this with simple black-box access of the anonymization model by constructing a pair-wise dataset of unanonymized and anonymized faces. To alleviate this issue, different defense measures that disrupts the traces in the anonymized image was investigated. The main take away from this, is that naively using what qualitatively looks convincing of hiding an identity is not necessary the case at all. Making robust quantitative evaluations important.

Previous studies have found that poor effort can significantly impact psychometric performance by Traumatic Brain Injury (TBI) patients. So far, this impact has been relatively well studied in attention and memory. However, this is not the case for visual perception functions. Thus, the goal of this study was to determine to what extent TBI severity affect visual perception after controlling for effort. Results showed that mild TBI good effort group did not differ from a demographically matched control group. In contrast, a mild TBI poor effort group, a moderate-severe TBI group and a right hemisphere cerebro-vascular (CVA) group performed worse than the mild TBI good effort group and the control group. The results suggest a dose response relationship between injury severity and visual perception performance. After controlling for effort, results indicated that moderate-severe TBI, but not mild TBI, has long lasting effects on visual perception. Clinical implications are discussed.

O reconhecimento de expressões faciais de emoções (REFE) possui valor adaptativo, sendo importante para o funcionamento social e o relacionamento interpessoal. Evidências apontaram que alterações no REFE podem estar associadas a transtornos psiquiátricos como a depressão maior (incluindo o subtipo psicótico) e ao estresse precoce. Associações também foram verificadas entre a vivência de estresse precoce e o desenvolvimento da depressão na vida adulta. Considerando esses aspectos, além da escassez de pesquisas na área envolvendo o subtipo psicótico da depressão, este estudo se justifica pela pertinência de averiguar se a interação entre a depressão psicótica e o estresse precoce altera o REFE. Assim, objetivou-se investigar a possível influência do episódio depressivo psicótico e da história de estresse precoce, considerados individualmente e em interação, sobre a acurácia e vieses no REFE. Adicionalmente, foi investigada a existência de associações entre a gravidade geral do estresse precoce e a acurácia específica no REFE. Participaram do estudo 49 pacientes com diagnóstico de primeiro episódio depressivo maior com características psicóticas e 49 controles saudáveis, emparelhados e subgrupados de acordo com o sexo e a história de estresse precoce. Todos os participantes passaram por uma bateria de avaliações, incluindo uma entrevista diagnóstica padronizada, escalas de avaliação da gravidade de sintomas psiquiátricos e funcionamento global, além de questionários sobre aspectos sociodemográficos, histórico clínico, uso de drogas e medicamentos e história de estresse precoce. Por fim, os participantes foram submetidos à Degraded Facial Affect Recognition Task (DFAR), uma tarefa comportamental de escolha forçada no REFE, composta por fotografias degradadas (i.e., com contraste visual reduzido em 30%) de expressões faciais de neutralidade, alegria, medo e raiva, às quais foram atribuídos rótulos emocionais correspondentes. Os dados foram analisados por meio de estatística descritiva e inferencial, utilizando os testes qui-quadrado ou exato de Fisher, U de Mann-Whitney, ANOVA de três vias com post hoc Bonferroni e Correlação de Spearman. Os principais resultados apontaram que a vivência de estresse precoce foi associada a uma maior acurácia no REFE de alegria nos controles e menor acurácia nos pacientes. O REFE de alegria também foi correlacionado com a gravidade geral do estresse precoce, tendo apresentado uma correlação positiva nos controles e negativa nos pacientes. Alterações no REFE de medo também foram associadas à interação entre depressão psicótica e estresse precoce, mas somente nas mulheres. Aquelas com diagnóstico de depressão psicótica e história de estresse precoce tiveram pior acurácia em comparação com mulheres depressivas psicóticas sem história de estresse precoce e com mulheres saudáveis com história de estresse precoce. Em conclusão, os resultados confirmaram que alterações no REFE foram associadas com a interação entre a depressão psicótica e o estresse precoce. Isso chama a atenção para a importância de se considerar a influência do estresse precoce em interação com outros transtornos psiquiátricos na investigação do reconhecimento emocional. / Facial emotion recognition (FER) has an adaptive value, being important for social functioning and interpersonal relationship. Evidence has indicated that FER alterations may be associated with psychiatric disorders such as major depression (including its psychotic subtype) and with early-life stress. Associations were also found between the experience of early-life stress and the development of depression in adult life. Considering these aspects, as well as the lack of research in this field involving the psychotic subtype of depression, this study is justified by the pertinence of investigating whether the interaction between psychotic depression and early-life stress alters FER. Thus, the objective was to investigate a possible influence of psychotic depressive episode and early-life stress history, considered individually and in interaction, on FER accuracy and bias. In addition, it was investigated the existence of associations between early-life stress general severity and specific FER accuracy. The study included 49 patients with diagnosis of first major depressive episode with psychotic features and 49 healthy controls, matched and subgrouped according to sex and earlylife stress history. All participants underwent a battery of assessments, including a standardized diagnostic interview, assessment scales on severity of psychiatric symptoms and global functioning, as well as questionnaires on sociodemographic aspects, clinical history, drug and medication use and early-life stress history. Finally, participants were submitted to the Degraded Facial Affect Recognition Task (DFAR), a FER forced-choice behavioral task, composed of degraded photographs (i.e., with visual contrast reduced by 30%) portraying facial expressions of neutrality, joy, fear and anger, to which a corresponding emotional label were assigned. Data were analyzed through descriptive and inferential statistics, using chi-square test or Fisher\'s exact test, Mann-Whitney U test, three-way ANOVA with Bonferroni post hoc test, and Spearman\'s correlation test. The main results pointed out that early-life stress experience was associated with greater accuracy in FER of joy in controls and lower accuracy in patients. FER of joy was also correlated with early-life stress general severity, presenting a positive correlation in controls and a negative correlation in patients. Alterations in FER of fear were also associated with the interaction between psychotic depression and early-life stress, but only for women. Those with psychotic depression diagnosis and early-life stress history had worse accuracy compared both to depressed psychotic women with no early-life stress history and to healthy women with early-life stress history. In conclusion, results confirmed that FER alterations were associated with the interaction between psychotic depression and early-life stress. This draws attention to the importance of considering the influence of early-life stress on interaction with other psychiatric disorders in the investigation of emotional recognition.

Há um crescente incentivo ao uso da tecnologia biométrica para melhorar, e até mesmo substituir os métodos tradicionais de segurança. O campo da Biometria refere-se a uma grande variedade de tecnologias usadas para identicar ou vericar a identidade de uma pessoa por meio da mensuração e análise de vários aspectos físicos e comportamentais do ser humano. Modalidades biométricas são características extraídas do corpo humano, que são únicas para cada indivíduo e que podem ser usadas para estabelecer sua identidade numa população. As principais modalidades biométricas empregadas na literatura são: impressão digital, face, voz, palma da mão e íris. Dentre as modalidades biométricas, a face é a mais comumente vista e usada em nossa vida diária. Em aplicações de mundo real, sistemas de reconhecimento facial, frequentemente, têm que lidar com condições não controladas e não previsíveis, tais como mudança na iluminação, pose, expressão e oclusão, as quais introduzem variações intraclasse e degradam a performance do reconhecimento. Comparada com problemas de pose, iluminação e expressão, o problema relacionado à oclusão é relativamente pouco estudado na área. As abordagens empregadas para atacar o problema de reconhecimento facial na presença de oclusões parciais podem ser categorizadas em: i) abordagens baseadas em reconstrução, que formulam o reconhecimento biométrico baseado em imagens de face parcialmente ocluídas como um problema de reconstrução; ii) abordagens baseadas em correspondência local, que extraem características locais da face; iii) abordagens baseadas em características insensíveis, que extraem características da face que são robustas à oclusão. O objetivo desta dissertação é realizar um estudo comparativo detalhado das abordagens baseadas em reconstrução visando reconhecimento biométrico facial na presença de oclusão. As abordagens baseadas em reconstrução foram aplicadas a duas bases de dados (AR e Yale) contendo quatro tipos diferentes de oclusões e avaliadas usando a performance obtida no reconhecimento biométrico. Com base no estudo comparativo realizado, foram propostas três técnicas baseadas em reconstrução, que correspondem a uma modicação e/ou combinação de técnicas já conhecidas na literatura. Os resultados experimentais mostraram que as técnicas propostas alcançaram desempenhos melhores ou similares as técnicas descritas na literatura / There is a growing incentive to use biometric technology to improve, even replace, traditional security methods. The eld of Biometrics refers to a wide variety of technologies used to identify or verify a persons identity by measuring and analyzing various physical and behavioral aspects of the human being. Biometric modalities are characteristics drawn from the human body, which are unique to each individual and can be used to establish their identity in a population. The main biometric modalities used in the literature are: fingerprint, face, voice, palm, and iris. The main biometric modalities used in the literature are: fingerprint, face, voice, palm and iris. Among the biometric modalities, the face is the most commonly seen and used in our daily life. In real-world applications, facial recognition systems often have to deal with uncontrolled and unpredictable conditions such as change in illumination, pose, expression, and occlusion, which introduce intraclass variations and degrade recognition performance. Compared with problems of pose, illumination and expression, the problem related to occlusion is relatively little studied in the area. The approaches used to tackle facial recognition in the presence of partial occlusions can be categorized into: (i) reconstruction-based approaches that formulate recognition of occluded faces as a reconstruction problem; ii) approaches based on local correspondence, which extract local face characteristics; iii) approaches based on insensitive characteristics, which extract features of the face that are robust to occlusion. The objective of this dissertation is to perform a detailed comparative study of reconstruction-based approaches aiming at facial recognition in the presence of occlusion. Reconstruction-based approaches were applied to two databases (AR and Yale) containing four dierent types of occlusions and evaluated using the performance obtained in the biometric recognition. Based on the comparative study, three reconstruction-based techniques were proposed, which correspond to a change in the formulation and/or combination of techniques already known in the literature. The experimental results showed that the proposed techniques achieved better or similar performances than the techniques described in the literature

Ce travail de thèse visait à concevoir, assembler, optimiser et tester des modules d’illumination structurée à base d’éléments optiques diffractifs (EODs) pour une application de reconnaissance faciale sur appareils mobiles (smartphones, tablettes). L’intégration des modules dans des smartphones impliquait de fortes contraintes de miniaturisation, de consommation énergétique, de coût, et de sécurité laser. L’élément clé de chaque module était un EOD de Fourier à angle de diffraction supérieur à la limite du modèle scalaire paraxial de la diffraction permettant d’illuminer la surface d’un visage à une distance d’une portée de bras. Afin de faciliter la conception (relâchement des contraintes angulaires), la fabrication (minimisation de l’efficacité de diffraction à l’ordre 0) et la réplication des EODs, le premier axe de travail a consisté à concevoir et à fabriquer des dispositifs hybrides « agrandisseurs d’angles » combinant des EODs et des optiques divergentes conventionnelles. Le second volet portait sur la conception des EODs qui devait prendre en considération à la fois les paramètres des systèmes bas coût d’illumination et d’acquisition d’images utilisés, notamment pour contrôler la présence de granularité laser (« speckle ») sur la figure de diffraction souhaitée (contrôle imposé par les algorithmes de reconnaissance faciale et de détection de fraudes utilisés). Le savoir-faire acquis dans le domaine de l’illumination structurée générée par des EODs a été étendu et transposé à trois autres applications dans les domaines de la vibrométrie, de l’aviation civile et commerciale, et de l’aviation militaire. / This thesis work aimed to design, assemble, optimize and test structured illumination modules based on diffractive optical elements (DOEs) for facial recognition application on mobile devices (smartphones, tablets). The integration of modules into smartphones involved significant constraints in terms of miniaturization, energy consumption, cost and laser safety. The key element of each module was a Fourier DOE with a diffraction angle greater than the limit of the paraxial scalar diffraction model to illuminate the surface of a face at a distance of an arm reach. In order to facilitate the design (relaxation of angular constraints), manufacturing (minimization of the zero order diffraction efficiency) and replication of DOEs, the first axis of research consisted in designing and manufacturing hybrid "angle enlarger" devices combining DOEs and conventional divergent optics. The second part concerned the design of the DOEs, which had to take into account both the parameters of the low-cost illumination and image acquisition systems used, in particular to control the presence of laser speckle on the desired diffraction pattern (control imposed by the facial recognition and fraud detection algorithms used). The know-how acquired in the field of structured illumination generated by DOEs has been extended and transposed to three other applications in the fields of vibrometry, civil and commercial aviation, and military aviation.

Identity verification systems are an important part of our every day life. A typical example is the Automatic Teller Machine (ATM) which employs a simple identity verification scheme: the user is asked to enter their secret password after inserting their ATM card; if the password matches the one prescribed to the card, the user is allowed access to their bank account. This scheme suffers from a major drawback: only the validity of the combination of a certain possession (the ATM card) and certain knowledge (the password) is verified. The ATM card can be lost or stolen, and the password can be compromised. Thus new verification methods have emerged, where the password has either been replaced by, or used in addition to, biometrics such as the person’s speech, face image or fingerprints. Apart from the ATM example described above, biometrics can be applied to other areas, such as telephone & internet based banking, airline reservations & check-in, as well as forensic work and law enforcement applications. Biometric systems based on face images and/or speech signals have been shown to be quite effective. However, their performance easily degrades in the presence of a mismatch between training and testing conditions. For speech based systems this is usually in the form of channel distortion and/or ambient noise; for face based systems it can be in the form of a change in the illumination direction. A system which uses more than one biometric at the same time is known as a multi-modal verification system; it is often comprised of several modality experts and a decision stage. Since a multi-modal system uses complimentary discriminative information, lower error rates can be achieved; moreover, such a system can also be more robust, since the contribution of the modality affected by environmental conditions can be decreased. This thesis makes several contributions aimed at increasing the robustness of single- and multi-modal verification systems. Some of the major contributions are listed below. The robustness of a speech based system to ambient noise is increased by using Maximum Auto-Correlation Value (MACV) features, which utilize information from the source part of the speech signal. A new facial feature extraction technique is proposed (termed DCT-mod2), which utilizes polynomial coefficients derived from 2D Discrete Cosine Transform (DCT) coefficients of spatially neighbouring blocks. The DCT-mod2 features are shown to be robust to an illumination direction change as well as being over 80 times quicker to compute than 2D Gabor wavelet derived features. The fragility of Principal Component Analysis (PCA) derived features to an illumination direction change is solved by introducing a pre-processing step utilizing the DCT-mod2 feature extraction. We show that the enhanced PCA technique retains all the positive aspects of traditional PCA (that is, robustness to compression artefacts and white Gaussian noise) while also being robust to the illumination direction change. Several new methods, for use in fusion of speech and face information under noisy conditions, are proposed; these include a weight adjustment procedure, which explicitly measures the quality of the speech signal, and a decision stage comprised of a structurally noise resistant piece-wise linear classifier, which attempts to minimize the effects of noisy conditions via structural constraints on the decision boundary.

Submitted by EDUARDO MACHADO SILVA (eduardodz@outlook.com) on 2018-06-02T22:50:24Z No. of bitstreams: 1 Eduardo_Final.pdf: 7020230 bytes, checksum: 17f5f419806417111d44cacbf46f3f0d (MD5) / Approved for entry into archive by Elza Mitiko Sato null (elzasato@ibilce.unesp.br) on 2018-06-04T16:05:04Z (GMT) No. of bitstreams: 1 silva_em_me_sjrp.pdf: 7020230 bytes, checksum: 17f5f419806417111d44cacbf46f3f0d (MD5) / Made available in DSpace on 2018-06-04T16:05:04Z (GMT). No. of bitstreams: 1 silva_em_me_sjrp.pdf: 7020230 bytes, checksum: 17f5f419806417111d44cacbf46f3f0d (MD5) Previous issue date: 2018-04-06 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O Reconhecimento facial é uma das tecnologias biométricas mais utilizadas em sistemas automatizados que necessitam garantir a identidade de uma pessoa para acesso autorizado e monitoramento. A grande aceitação do uso da face tem várias vantagens sobre outras tecnologias biométricas: ela é natural, não exige equipamentos sofisticados, a aquisição de dados é baseada em abordagens não invasivas, e pode ser feito a distância, de maneira cooperativa ou não. Embora muitos estudos em reconhecimento facial tenham sido feitos, problemas com variação de iluminação, poses com oclusão facial, expressão facial e envelhecimento ainda são desafios, pois influenciam a performance dos sistemas de reconhecimento facial e motivam o desenvolvimento de novos sistemas de reconhecimento que lidam com esses problemas e sejam mais confiáveis. Este trabalho tem como objetivo avaliar a técnica de Padrões Localmente Mapeados em Multiescala (MSLMP) para o reconhecimento facial. Técnicas baseadas em algoritmos genéticos e processamento de imagens foram usadas para obter melhores resultados. Os resultados obtidos chegam a 100% de acurácia para alguns banco de dados. A base de dados MUCT ´e, em particular, bastante complexa, ela foi criada em 2010 com o objetivo de aumentar a quantidade de bancos de dados disponíveis com alta variação de iluminação, idade, posições e etnias, e por isso, ´e um banco de dados difícil quanto ao reconhecimento automático de faces. Uma nova técnica de processamento baseada na média dos níveis de cinza da base foi desenvolvida. / Facial recognition is one of the most used biometric technologies in automated systems which ensure a person’s identity for authorized access and monitoring. The acceptance of face use has several advantages over other biometric technologies: it is natural, it does not require sophisticated equipment, data acquisition is based on non-invasive approaches, and can it be done remotely, cooperatively or not. Although many facial recognition studies have been done, problems with light variation, facial occlusion, position, expression, and aging are still challenges, because they influence the performance of facial recognition systems and motivate the development of more reliable recognition systems that deal with these problems. This work aim to evaluate the Multi-scale Local Mapped Pattern (MSLMP) technique for the facial recognition. Techniques based on genetic algorithms and image processing were applied to increase the performance of the method. The obtained results reach up to 100% of accuracy for some databases. A very difficult database to deal is the MUCT database which was created in 2010 with aim of providing images with high variation of lighting, age, positions and ethnicities in the facial biometry literature, which makes it a highly difficult base in relation to automated recognition. A new processing technique was developed based on the average gray levels of the images of the database.

Orientador: Aparecido Nilceu Marana / Banca: Hélio Pedrini / Banca: Aledir Silveira Pereira / Resumo: As técnicas para identificação biométrica têm evoluído cada vez mais devido à necessidade que os seres humanos têm de identificar as pessoas em tempo real e de forma precisa para permitir o acesso a determinados recursos, como por exemplo, as aplicações e serviços WEB. O reconhecimento facial é uma técnica biométrica que apresenta várias vantagens em relação às demais, tais como: uso de equipamentos simples e baratos para a obtenção das amostras e a possibilidade de se realizar o reconhecimento em sigilo e à distância. O reconhecimento de faces a partir de vídeo é uma tendência recente na área de Biometria. Esta dissertação tem por objetivo principal comparar diferentes técnicas de reconhecimento facial a partir de vídeo para determinar as que apresentam um melhor compromisso entre tempo de processamento e precisão. Outro objetivo é a incorporação dessas melhores técnicas no sistema de autenticação biométrica em ambientes de E-Learning, proposto em um trabalho anterior. Foi comparado o classificador vizinho mais próximo usando as medidas de distância Euclidiana e Mahalanobis com os seguintes classificadores: Redes Neurais MLP e SOM, K Vizinhos mais Próximos, Classificador Bayesiano, Máquinas de Vetores de Suporte (SVM) e Floresta de Caminhos Ótimos (OPF). Também foi avaliada a técnica de Modelos Ocultos de Markov (HMM). Nos experimentos realizados com a base Recogna Video Database, criada especialmente para uso neste trabalho, e Honda/UCSD Video Database, os classificadores apresentaram os melhores resultados em termos de precisão, com destaque para o classificador SVM da biblioteca SVM Torch. A técnica HMM, que incorpora informações temporais, apresentou resultados melhores do que as funções de distância, em termos de precisão, mas inferiores aos classificadores / Abstract: The biometric identification techniques have evolved increasingly due to the need that humans have to identify people in real time to allow access to certain resources, such as applications and Web services. Facial recognition is a biometric technique that has several advantages over others. Some of these advantages are the use of simple and cheap equipment to obtain the samples and the ability to perform the recognition in covert mode. The face recognition from video is a recent approach in the area of Biometrics. The work in this dissertation aims at comparing different techniques for face recognition from video in order to find the best rates on processing time and accuracy. Another goal is the incorporation of these techniques in the biometric authentication system for E-Learning environments, proposed in an earlier work. We have compared the nearest neighbor classifier using the Euclidean and Mahalanobis distance measures with some other classifiers, such as neural networks (MLP and SOM), k-nearest neighbor, Bayesian classifier, Support Vector Machines (SVM), and Optimum Path Forest (OPF). We have also evaluated the Hidden Markov Model (HMM) approach, as a way of using the temporal information. In the experiments with Recogna Video Database, created especially for this study, and Honda/UCSD Video Database, the classifiers obtained the best accuracy, especially the SVM classifier from the SVM Torch library. HMM, which takes into account temporal information, presented better performance than the distance metrics, but worse than the classifiers / Mestre

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2017-05-05T13:56:17Z No. of bitstreams: 2 Dissertação - Cleiton Paiva Aquino - 2017.pdf: 1798785 bytes, checksum: 24ba3dcb878f81a76626a573db9c0a44 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-05-05T13:56:38Z (GMT) No. of bitstreams: 2 Dissertação - Cleiton Paiva Aquino - 2017.pdf: 1798785 bytes, checksum: 24ba3dcb878f81a76626a573db9c0a44 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-05-05T13:56:38Z (GMT). No. of bitstreams: 2 Dissertação - Cleiton Paiva Aquino - 2017.pdf: 1798785 bytes, checksum: 24ba3dcb878f81a76626a573db9c0a44 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-03-04 / Facial aging is known as a complex process that varies the shape and texture of thefacial area. Variations in shape include variation in craniofacial structure, while texture variation includes skin coloration, appearance of facial lines and wrinkles. Shape and texture are both considered the most common forms of facial aging patterns. Fiducial points are control points on an object that define characteristic regions with properties of interest to the application. Thus, the objective of this work is to evaluate the behavior offacial fiducial points during human aging. An evaluation was performed in a statistical manner and through classification, through the characteristics vectors obtained throughfacial fiducial points. For the accomplishment of this work, we had a social motivation,that is the aid in the search for missing persons, and another one of technical character,that is the perfection of facial recognition systems. Among the results we can highlight the behavior of increase, reduction and stabilization among some fiducial points. With regard to classification, we obtained a result of 84.29 % of correct answers when we compared the class with people under 20 years old and the class with people between 20 and 39 years old from the black men’s groups / O envelhecimento facial é conhecido como um complexo processo que varia a forma ea textura da área facial. Variações na forma incluem variação na estrutura craniofacial, enquanto a variação na textura inclui coloração da pele, aparecimento de linhas faciais e rugas. A forma e a textura são ambas consideradas as formas mais comuns de padrões de envelhecimento facial. Já os pontos fiduciais são pontos de controle sobre um objeto que definem regiões características com propriedades de interesse à aplicação. Assim, o objetivo deste trabalho é avaliar o comportamento dos pontos fiduciais faciais durante o envelhecimento humano. Foi realizada uma avaliação de forma estatística e por meio de classificação, através dos vetores de características obtidos através dos pontos fiduciaisfaciais. Para a realização deste trabalho, tivemos uma motivação de cunho social, queé o auxílio na busca por pessoas desaparecidas, e outra de cunho técnico, que é o aperfeiçoamento de sistemas de reconhecimento facial. Dentre os resultados podemos salientar o comportamento de aumento, redução e estabilização entre alguns pontos fiduciais. No tocante a classificação chegamos a obter um resultado de 84,29% de acertos quando comparamos a classe com pessoas menores de 20 anos e a classe com pessoas entre 20 e 39 anos do grupos de homens negros.

En la actualidad, el incremento de actos delictivos se ha hecho más frecuente en Lima, por lo que es considerado como uno de los mayores problemas que afecta a nuestra sociedad. Esto se ve reflejado en los últimos reportes de seguridad ciudadana donde Lima es considerada una de las ciudades más peligrosas de América Latina. A su vez, las autoridades no cuentan con las herramientas suficientes para combatir a la delincuencia generando que la población se sienta insegura de caminar por las calles y aumentando la cantidad de actos delictivos. Esto incluye la escasez de tecnología que permita optimizar los procesos de identificación o captura de delincuentes, ya que existe un lapso entre que el acto delictivo se realizó y la llegada de las autoridades a la escena. Este tiempo es de suma importancia para poder capturar a un delincuente o poder identificarlo. El presente proyecto tiene como objetivo brindar una solución tecnológica para las autoridades correspondientes que permita optimizar el tiempo de identificación de delincuentes a través del uso del reconocimiento facial. Esta tecnología permite reconocer a delincuentes mediante un análisis de los rasgos faciales del sospechoso y poder realizar una comparación con la base de datos de los delincuentes. La solución tecnológica facultara una alerta en tiempo real con los datos del sospechoso identificado y la ubicación de la cámara que lo idéntico. Esto permitirá que las autoridades puedan realizar mejores estrategias logrando anticiparse al hecho delictivo o capturando al criminal. Los resultados preliminares mostraron que la solución tecnológica permite reducir el tiempo actual de identificación de 5 minutos aproximadamente en un 91.46% y cuenta con una eficacia del 90% basándonos en las pruebas realizadas que son detallas en los capítulos correspondientes / At present, the increase in criminal acts has become more frequent in Lima. For this reason, it is considered one of the biggest problems affecting our society. We can see it in the latest citizen security reports, where Lima is one of the most dangerous cities in Latin America. Another fact is that the authorities do not have enough tools to fight crime, making the population feel unsafe to walk the streets and increasing the number of criminal acts. This problem includes the scarcity of technology that allows optimizing the processes of identification or capture of criminals since there is a period between when the criminal act was carried out and the arrival of the authorities on the scene. This time is of the utmost importance to be able to capture a criminal or to be able to identify him. The objective of this project is to provide a technological solution for the corresponding authorities that allows optimizing the identification time of criminals through the use of facial recognition. This technology makes it possible to recognize criminals through an analysis of the suspect's facial features and to be able to make a comparison with the criminal database. The technological solution will enable an alert in real-time with the data of the identified suspect and the location of the identical camera. This will allow the authorities to carry out better strategies, anticipating the crime, or capturing the criminal. Preliminary results showed that the technological solution reduces the current identification time of 5 minutes approximately by 91.46% and has an efficiency of 90% based on the tests carried out, which are detailed in the corresponding chapters. / Tesis