It is estimated that there are currently fifty million people living with dementia worldwide. An accurate and early diagnosis of dementia is important in order to initiate appropriate treatment programs as soon as possible. Common methods of neuropsychological assessment can be sensitive to external factors which may compromise accuracy. The aim of this thesis was to investigate techniques that have the potential for the detection of dementia that avoid some of the external influences. The thesis looked at measurements of (i) postural stability (ii) facial analysis and (iii) fully-immersive virtual reality in cognitively-healthy individuals. These techniques were chosen as postural stability and facial analysis change in dementia and whilst virtual reality has previously been used in dementia research, fully-immersive virtual reality measures have not been established. To see if the measurements were associated with cognitive function, participants completed a series of cognitive tests. Results indicate that all techniques explored shared a relationship with memory performance, with lower anteroposterior postural sway (F(1,22) = 17.76, p < 0.01), number of activities participated in that involve a posture element (F(2, 39) = 3.77, p < .05; Wilk's Λ = 0.84, partial η2 = 0.16), the greater the frequency of negative facial expressions (F(2, 18) = 4.49, p < .05; Wilk's Λ = 0.67, partial η2 = 0.33.), and low blink rate (t(11.02) = 2.62 p < .05) all showing better scores on memory tests. Moreover, better scores on the fullyimmersive virtual reality task predicted better scores on with short-term memory (F(1,22) = 20.20 p < 0.01), LTM (F(1,22) = 09.10 p < .01), associative learning (F(1,22) = 08.75 p < .01), and a dual–task test (F(1,22) = 04.64 p < .05). The novel findings that elements such as postural stability, participation in sports, facial expressions of emotion, blink rates, and spatial memory as assessed in fully-immersive virtual reality highlight that non-invasive techniques can provide measurements that correspond to cognitive ability. This may hold implications for dementia diagnoses. Future research should assess whether these relationships can also be found in an older adult population. If this relationship is found in older adults, it could justify further research into how these techniques could be applied in a clinical context.

Yes / Emerging applications in surveillance, the entertainment industry and other human computer interaction applications have motivated the development of real-time facial analysis research covering detection, tracking and recognition. In this paper, the authors present a review of recent facial analysis for real-time applications, by providing an up-to-date review of research efforts in human computing techniques in the visible domain. The main goal is to provide a comprehensive reference source for researchers, regardless of specific research areas, involved in real-time facial analysis. First, the authors undertake a thorough survey and comparison in face detection techniques. In this survey, they discuss some prominent face detection methods presented in the literature. The performance of the techniques is evaluated by using benchmark databases. Subsequently, the authors provide an overview of the state-of-the-art of facial expressions analysis and the importance of psychology inherent in facial expression analysis. During the last decades, facial expressions analysis has slowly evolved into automatic facial expressions analysis due to the popularity of digital media and the maturity of computer vision. Hence, the authors review some existing automatic facial expressions analysis techniques. Finally, the authors provide an exemplar for the development of a facial analysis real-time application and propose a model for facial analysis. This review shows that facial analysis for real-time application involves multi-disciplinary aspects and it is important to take all domains into account when building a reliable system.

A face é a parte do corpo que mais sintetiza o ser humano e, para que seja entendida, vem sendo alvo de estudos dos mais diversos campos da ciência, incluindo a área da Identificação Facial Forense. A ciência responsável pelo estudo da face e de suas relações, denominada Cefalometria, apesar de exaustivamente estudada para o emprego clínico e radiográfico, principalmente por profissionais da Odontologia, ainda não foi padronizada e validada para o seu emprego em análises faciais sobre imagens e/ou fotografias. Em virtude do crescimento da produção tecnológica e automatizada da era em que vivemos, onde aparelhos com câmeras fotográficas acopladas fazem parte da vida corriqueira da grande maioria da população mundial, é necessário o desenvolvimento de metodologias que confirmem a objetividade, reprodutibilidade e confiabilidade de suas análises, principalmente para as áreas que exigem precisão e rigor técnico-científico, como a perícia criminal. Neste sentido, o presente trabalho teve como objetivo propor uma metodologia de normatização na determinação de pontos cefalométricos para exames faciais exclusivamente baseados em imagens em norma frontal e definir quais são os pontos cefalométricos que apresentam maior e menor variabilidade de aferição, para possível aplicação na identificação humana. A parte experimental do estudo foi delineada em duas etapas: na primeira fase, foi utilizado o método convencional (cefalométrico clássico) para a descrição dos pontos anatômicos de referência; na segunda, o método fotoantropométrico proposto. Ambas as fases foram analisadas por cinco examinadores, os quais marcaram 16 pontos na topografia facial de 18 imagens, aleatoriamente escolhidas de um banco de imagens. Para todas as análises, foi utilizado um software desenvolvido pelo Serviço de 13 Perícias em Audiovisual e Eletrônicos do Instituto Nacional de Criminalística da Polícia Federal, denominado de SMVFace. Pode-se observar que houve uma grande redução da variabilidade dos pontos anatômicos após a adoção da descrição fotoantropométrica, principalmente para os pontos Alar, Endocanthion, Glabela, Gônio, Irídio Medial, Labial Superior e Zígio. Contrariamente, um ligeiro aumento na variabilidade dos pontos Chelion, Labial Inferior e Gnátio foi observado. Apesar do aumento na variabilidade desses pontos, a maioria das marcações foi realizada abaixo do erro aceitável preconizado por alguns estudos. Esses resultados demonstraram que, apesar de algumas limitações, a utilização da descrição fotoantropométrica proposta foi determinante para alcançar uma maior precisão das marcações dos pontos anatômicos de referência de forma geral / Face is the body part that most synthesizes the human being and, to be understood, has been targeted from different branches of science, including the Forensic Facial Identification area. The Cephalometry, science responsible for facial studies and their relationships, although extensively studied for clinical and radiographic employment, especially by dental professionals, it has not been standardized and validated yet for its use in facial images and/or photo analysis. Due to the growth of technology and automated production of the era in which we live, where camera-coupled devices are part of vast majority ordinary life of world population, methodologies development is necessary to confirm the objectivity, reproducibility and reliability of their analyzes, especially for areas that require precision and technical/scientific rigor as criminal expertise. In this sense, this research aimed to propose a standardization method in determining cephalometric points exclusively based on frontal view facial images and determine which points have higher and lower variability of measurement, for possible use in human identification. The experimental part of the study was drawn in two stages. In the first phase, the conventional method (classic cephalometry) was used to describe the reference anatomical points and, in the second, the proposed photoanthropometric method was used. Both phases were analyzed by five examiners who scored 16 points in facial topography of 18 images randomly chosen from an image database. For all analyzes, a software developed by the Expertise Service in Audiovisual and Electronics of the National Institute of Criminology, Federal Police, called SMVFace, was used. It could be observed that there was a large variability reduction of anatomical points after the photoanthropometry description adoption, especially for 15 Alar, Endocanthion, Glabella, Gonion, Iridium Medial, Upper Lip and Zigion points. In contrast, a slight variability increase of Chelion, Lower Lip and Gnathion points was observed. Despite the increased variability of these points, most markings are performed below the acceptable error advocated by some studies. These results demonstrate that, although some limitations, the use of the proposed photoanthropometric description was crucial to achieve greater accuracy of anatomical landmarks determination in general

A face é a parte do corpo que mais sintetiza o ser humano e, para que seja entendida, vem sendo alvo de estudos dos mais diversos campos da ciência, incluindo a área da Identificação Facial Forense. A ciência responsável pelo estudo da face e de suas relações, denominada Cefalometria, apesar de exaustivamente estudada para o emprego clínico e radiográfico, principalmente por profissionais da Odontologia, ainda não foi padronizada e validada para o seu emprego em análises faciais sobre imagens e/ou fotografias. Em virtude do crescimento da produção tecnológica e automatizada da era em que vivemos, onde aparelhos com câmeras fotográficas acopladas fazem parte da vida corriqueira da grande maioria da população mundial, é necessário o desenvolvimento de metodologias que confirmem a objetividade, reprodutibilidade e confiabilidade de suas análises, principalmente para as áreas que exigem precisão e rigor técnico-científico, como a perícia criminal. Neste sentido, o presente trabalho teve como objetivo propor uma metodologia de normatização na determinação de pontos cefalométricos para exames faciais exclusivamente baseados em imagens em norma frontal e definir quais são os pontos cefalométricos que apresentam maior e menor variabilidade de aferição, para possível aplicação na identificação humana. A parte experimental do estudo foi delineada em duas etapas: na primeira fase, foi utilizado o método convencional (cefalométrico clássico) para a descrição dos pontos anatômicos de referência; na segunda, o método fotoantropométrico proposto. Ambas as fases foram analisadas por cinco examinadores, os quais marcaram 16 pontos na topografia facial de 18 imagens, aleatoriamente escolhidas de um banco de imagens. Para todas as análises, foi utilizado um software desenvolvido pelo Serviço de 13 Perícias em Audiovisual e Eletrônicos do Instituto Nacional de Criminalística da Polícia Federal, denominado de SMVFace. Pode-se observar que houve uma grande redução da variabilidade dos pontos anatômicos após a adoção da descrição fotoantropométrica, principalmente para os pontos Alar, Endocanthion, Glabela, Gônio, Irídio Medial, Labial Superior e Zígio. Contrariamente, um ligeiro aumento na variabilidade dos pontos Chelion, Labial Inferior e Gnátio foi observado. Apesar do aumento na variabilidade desses pontos, a maioria das marcações foi realizada abaixo do erro aceitável preconizado por alguns estudos. Esses resultados demonstraram que, apesar de algumas limitações, a utilização da descrição fotoantropométrica proposta foi determinante para alcançar uma maior precisão das marcações dos pontos anatômicos de referência de forma geral / Face is the body part that most synthesizes the human being and, to be understood, has been targeted from different branches of science, including the Forensic Facial Identification area. The Cephalometry, science responsible for facial studies and their relationships, although extensively studied for clinical and radiographic employment, especially by dental professionals, it has not been standardized and validated yet for its use in facial images and/or photo analysis. Due to the growth of technology and automated production of the era in which we live, where camera-coupled devices are part of vast majority ordinary life of world population, methodologies development is necessary to confirm the objectivity, reproducibility and reliability of their analyzes, especially for areas that require precision and technical/scientific rigor as criminal expertise. In this sense, this research aimed to propose a standardization method in determining cephalometric points exclusively based on frontal view facial images and determine which points have higher and lower variability of measurement, for possible use in human identification. The experimental part of the study was drawn in two stages. In the first phase, the conventional method (classic cephalometry) was used to describe the reference anatomical points and, in the second, the proposed photoanthropometric method was used. Both phases were analyzed by five examiners who scored 16 points in facial topography of 18 images randomly chosen from an image database. For all analyzes, a software developed by the Expertise Service in Audiovisual and Electronics of the National Institute of Criminology, Federal Police, called SMVFace, was used. It could be observed that there was a large variability reduction of anatomical points after the photoanthropometry description adoption, especially for 15 Alar, Endocanthion, Glabella, Gonion, Iridium Medial, Upper Lip and Zigion points. In contrast, a slight variability increase of Chelion, Lower Lip and Gnathion points was observed. Despite the increased variability of these points, most markings are performed below the acceptable error advocated by some studies. These results demonstrate that, although some limitations, the use of the proposed photoanthropometric description was crucial to achieve greater accuracy of anatomical landmarks determination in general

Nesta pesquisa foi comparada a diferença das medidas obtidas utilizando-se a análise cefalométrica do perfil tegumentar tendo como base a análise facial de Arnett relacionada à Horizontal Verdadeira com a análise cefalométrica do perfil tegumentar feita a partir do Plano de Frankfurt. Utilizamos 140 radiografias cefalométricas requisitadas como parte de documentação ortodôntica, obtidas de pacientes com idade variando de 06 a 49 anos, sendo 64 do sexo masculino e 76 do sexo feminino, pertencentes ao arquivo do IOM - (Instituto de Odontologia Multidisciplinar) ? RJ. As radiografias cefalométricas em norma lateral foram obtidas pelo mesmo operador, obedecendo ao protocolo adotado pelo IOM. Todos os pacientes foram orientados a assumir a Posição Natural da Cabeça, tendo como referência um espelho posicionado a frente e a Linha Vertical Verdadeira foi obtida por meio da utilização de um fio metálico unido à um prumo de chumbo, posicionado próximo à margem anterior do chassi porta-filme, de forma que apareça à frente do perfil tegumentar do paciente. A partir das medidas lineares obtidas nas Análises Cefalométricas dos Tecidos Moles, foram realizadas análises estatísticas visando dois objetivos: comparar as medidas de A à I obtidas nos traçados (um utilizando como referência uma perpendicular ao Plano Horizontal de Frankfurt e o outro à Linha Vertical Verdadeira) e analisar os resultados comparativamente. Após a análise estatística (Kolmogorov-Smirnov, para distribuição normal de dados; t-Student para amostras emparelhadas e teste não paramétrico de Wilcoxon) e a interpretação dos resultados obtidos, concluímos que houve diferenças significantes entre as formas de medidas das variáveis A, B, C, D, F, G e I, pois o valor do nível descritivo (p-valor) se apresentou abaixo de 0,05. Entretanto para as medidas E e H o valor do nível descritivo (p-valor) se apresentou maior que 0,05, demonstrando não haver diferenças significantes entre as formas dessas medidas. As correlações entre as medidas apresentadas estabelecidas pela análise facial de Arnett, de acordo com este estudo, não deverão ser utilizadas em pacientes que se submeteram à radiografias cefalométricas realizadas a partir do Plano Horizontal de Frankfurt, salvo no caso da distância interlabial, representada pela medida H, que além de não apresentar diferenças significantes neste estudo, não tem, segundo a análise de Arnett, correlação com nenhuma outra medida. / In this research a comparison was made between the differences in measurements obtained using the soft tissue profile?s cephalometric analysis based on Arnett?s facial analysis related to the true vertical with the cephalometric analysis of soft tissue profile based on the Frankfurt Plan. A total of 140 lateral cephalometric radiographs were used as part of the orthodontic documentation, taken from patients whose ages ranged from 06 to 49 years of age, of these 64 males and 76 females, being part of the files of the IOM (Instituto de Odontologia Multidisciplinar do Rio de Janeiro). The lateral cephalometric radiographs were obtained from the same operator obeying the protocols adapted by the IOM. All patients were oriented to place themselves in the Natural Head Posture using, as a reference, a mirror in front of then and the true vertical line was obtained using a metal wire and a plumb-line, placed close to the frontal chassis of the film chamber, in such a way that it appears in front of the soft tissue profile of the patient. Parting from the lineal measurements obtained from the Cephalometric Analyses of the Soft Tissues statistical analyses were made relating to two objectives: to compare the measurements from A to I obtained from the drafts/ sketches (one using as it reference a perpendicular to the Frankfurt Horizontal Plane and the other the True Vertical Line), comparatively analyse these results. After the statistical analysis (Kolgomorov Smirnov) for a normal data distribution: t-student for matched pairs and the no-parametric Wilcanox test) and the analysis of the results obtained, it was concluded that there were significant differences between the forms of the measurements of the variables A, B, C, D, F, G and I since the descriptive value (p-value) was below 0,05. The measurements for E and H, however, being over 0,05, showing than there were no significant differences between these forms of measurement. The correlation between the demonstrated measurements established by Arnett?s facial analysis, in accordance with this study, should not be used in patients who have undergone lateral cephalometric radiograph initiated on the basis of the Frankfurt Horizontal Plane except in the case of interlabial distance represented by the measurement H, which even differences in this study, does not have, according to Arnett?s studies, correlations with any other measurements.

Made available in DSpace on 2016-08-03T16:31:06Z (GMT). No. of bitstreams: 1 Dissertacao Olivia Morihisa.pdf: 1468336 bytes, checksum: 2be8ff54a9d3afe65cb3dc995e2fba72 (MD5) Previous issue date: 2006-03-30 / The purpose of this research was to study some facial analysis used in orthodontic diagnosis and to verify the agreement between lateral and frontal views in the evaluation of the face attractiveness done by lay and professional groups, the agreement between these groups in the evaluation of the face attractiveness on both lateral and frontal views, as well as to verify the association between face attractiveness and Facial Golden Proportion, face attractiveness and Facial Pattern and between Facial Pattern and Facial Golden Proportion. Two hundred and eight standardized facial photographs (104 in lateral view and 104 in frontal view) of a hundred and four randomly chosen individuals were used. The pictures were first classified in pleasant , acceptable and not pleasant by the Orthodontics and Layman groups. Then Facial Golden Proportion measurements were made in lateral and frontal views by means of a computerized program and the individuals were classified on their Facial Pattern by its lateral aspect. After statistical analysis, it was noted that there was no agreement between groups and views in the facial attractiveness classification, as well as in the association between Golden Proportion measurements and facial attractiveness or Facial Pattern. Considering Facial Pattern and attractiveness, there was strong association in the lateral view, however in the frontal view there was no association, for both observers groups. Key words: Orthodontics; diagnosis; Facial analysis; photograph; Golden Proportion. / O propósito desta pesquisa foi estudar algumas análises faciais utilizadas para diagnóstico ortodôntico e verificar a concordância entre norma lateral e frontal na avaliação da agradabilidade facial para os grupos leigos e profissionais, a concordância entre estes grupos na avaliação da agradabilidade facial nas normas lateral e frontal, bem como verificar a associação entre agradabilidade facial e Proporção Áurea, agradabilidade facial e Padrão Facial e entre Padrão Facial e Proporção Áurea. Utilizou-se 208 fotografias faciais padronizadas (104 laterais e 104 frontais) de 104 indivíduos escolhidos aleatoriamente, que primeiramente foram classificadas em agradável , aceitável e desagradável por dois grupos distintos: grupo Ortodontia e grupo Leigos . As fotografias laterais e frontais foram submetidas a medidas de Proporção Áurea Facial por meio de programa computadorizado e os indivíduos foram classificados quanto ao Padrão Facial pelo seu aspecto lateral. Após análise estatística, verificou-se que não houve concordância entre as variáveis da avaliação de agradabilidade estudadas, bem como não houve associação entre Proporção Áurea com agradabilidade facial ou com Padrão Facial. Entre agradabilidade facial e Padrão Facial, observou-se para a norma lateral associação fortemente positiva, porém para a frontal não houve associação para ambos os grupos de avaliadores.

Made available in DSpace on 2016-08-03T16:31:11Z (GMT). No. of bitstreams: 1 APENDICE.pdf: 12905 bytes, checksum: a8c8bd261256029ebcec43c5ee7829d2 (MD5) Previous issue date: 2008-02-26 / The objective of this research was to evaluate the facial alterations of surgically assisted rapid maxillary expansion (SARME). The sample consisted of 15 patients with 24,1 years old (average), 10 female and 5 male, that shown transverse discrepancy of the maxillary arch, the patients didn t have any orthodontic treatment before, they had complete clinical documentation and photographs in frontal view in the pre-treatment step (T1) and 6 months after the (SARME) (T2). Linear measurements were taken by reference points marked in acetate sheets fixed over the photographs, to avoid the anatomic drawing execution. The conclusion was that it is extremely important to standardize the photographs in all treatment phases to trust the obtained measurements. When T1 and T2 were compared by Student t test, there weren t any significant statistics alterations in: intercantal distance (Ind Ine); medium facial length (N - SN), right eye width (Exd Ind), left eye width (Exe Ine); facial length (N Me ), Upper face width (Zid Zie ), mouth width (Cbd Cbe) and mouth length (Ls Li). Inferior facial length (Sn - Me ) and noise width (Ald Ale) measurements showed significant statistics alterations after (SARME). / O objetivo desta pesquisa foi avaliar as alterações faciais decorrentes da Expansão Rápida da Maxila Assistida Cirurgicamente (ERM-AC). A amostra foi composta por 15 pacientes com idade média de 24 anos e 1 mês, sendo 10 do sexo feminino e 5 do sexo masculino, que apresentavam deficiência transversal da maxila, não tinham sido submetidos a tratamento ortodôntico prévio, apresentavam ficha clínica completa e fotografias em norma frontal nas fases pré-tratamento (T1) e 6 meses após a ERM-AC (T2). Mediadas lineares foram obtidas a partir da marcação de pontos de referência em folhas de acetato fixadas sobre as fotografias, para evitar a necessidade de execução de desenho anatômico. Concluiu-se que a padronização de fotografias em todos os tempos da pesquisa é de fundamental importância para que as medidas avaliadas sejam confiáveis. Quando comparados T1 com T2 por meio do teste t de Student não se verificou alteração estatisticamente significante na: largura intercantal (Ind Ine), altura facial média (N - SN), largura do olho direito (Exd Ind), largura do olho esquerdo (Exe Ine), altura facial (N - Me ), largura facial superior (Zid - Zie ), largura da boca (Cbd Cbe) e altura da boca (Ls Li). As medidas altura facial inferior (Sn - Me ) e a largura do nariz (Ald Ale) apresentaram alteração estatisticamente significante após a ERM-AC.

Nesta pesquisa foi comparada a diferença das medidas obtidas utilizando-se a análise cefalométrica do perfil tegumentar tendo como base a análise facial de Arnett relacionada à Horizontal Verdadeira com a análise cefalométrica do perfil tegumentar feita a partir do Plano de Frankfurt. Utilizamos 140 radiografias cefalométricas requisitadas como parte de documentação ortodôntica, obtidas de pacientes com idade variando de 06 a 49 anos, sendo 64 do sexo masculino e 76 do sexo feminino, pertencentes ao arquivo do IOM - (Instituto de Odontologia Multidisciplinar) ? RJ. As radiografias cefalométricas em norma lateral foram obtidas pelo mesmo operador, obedecendo ao protocolo adotado pelo IOM. Todos os pacientes foram orientados a assumir a Posição Natural da Cabeça, tendo como referência um espelho posicionado a frente e a Linha Vertical Verdadeira foi obtida por meio da utilização de um fio metálico unido à um prumo de chumbo, posicionado próximo à margem anterior do chassi porta-filme, de forma que apareça à frente do perfil tegumentar do paciente. A partir das medidas lineares obtidas nas Análises Cefalométricas dos Tecidos Moles, foram realizadas análises estatísticas visando dois objetivos: comparar as medidas de A à I obtidas nos traçados (um utilizando como referência uma perpendicular ao Plano Horizontal de Frankfurt e o outro à Linha Vertical Verdadeira) e analisar os resultados comparativamente. Após a análise estatística (Kolmogorov-Smirnov, para distribuição normal de dados; t-Student para amostras emparelhadas e teste não paramétrico de Wilcoxon) e a interpretação dos resultados obtidos, concluímos que houve diferenças significantes entre as formas de medidas das variáveis A, B, C, D, F, G e I, pois o valor do nível descritivo (p-valor) se apresentou abaixo de 0,05. Entretanto para as medidas E e H o valor do nível descritivo (p-valor) se apresentou maior que 0,05, demonstrando não haver diferenças significantes entre as formas dessas medidas. As correlações entre as medidas apresentadas estabelecidas pela análise facial de Arnett, de acordo com este estudo, não deverão ser utilizadas em pacientes que se submeteram à radiografias cefalométricas realizadas a partir do Plano Horizontal de Frankfurt, salvo no caso da distância interlabial, representada pela medida H, que além de não apresentar diferenças significantes neste estudo, não tem, segundo a análise de Arnett, correlação com nenhuma outra medida. / In this research a comparison was made between the differences in measurements obtained using the soft tissue profile?s cephalometric analysis based on Arnett?s facial analysis related to the true vertical with the cephalometric analysis of soft tissue profile based on the Frankfurt Plan. A total of 140 lateral cephalometric radiographs were used as part of the orthodontic documentation, taken from patients whose ages ranged from 06 to 49 years of age, of these 64 males and 76 females, being part of the files of the IOM (Instituto de Odontologia Multidisciplinar do Rio de Janeiro). The lateral cephalometric radiographs were obtained from the same operator obeying the protocols adapted by the IOM. All patients were oriented to place themselves in the Natural Head Posture using, as a reference, a mirror in front of then and the true vertical line was obtained using a metal wire and a plumb-line, placed close to the frontal chassis of the film chamber, in such a way that it appears in front of the soft tissue profile of the patient. Parting from the lineal measurements obtained from the Cephalometric Analyses of the Soft Tissues statistical analyses were made relating to two objectives: to compare the measurements from A to I obtained from the drafts/ sketches (one using as it reference a perpendicular to the Frankfurt Horizontal Plane and the other the True Vertical Line), comparatively analyse these results. After the statistical analysis (Kolgomorov Smirnov) for a normal data distribution: t-student for matched pairs and the no-parametric Wilcanox test) and the analysis of the results obtained, it was concluded that there were significant differences between the forms of the measurements of the variables A, B, C, D, F, G and I since the descriptive value (p-value) was below 0,05. The measurements for E and H, however, being over 0,05, showing than there were no significant differences between these forms of measurement. The correlation between the demonstrated measurements established by Arnett?s facial analysis, in accordance with this study, should not be used in patients who have undergone lateral cephalometric radiograph initiated on the basis of the Frankfurt Horizontal Plane except in the case of interlabial distance represented by the measurement H, which even differences in this study, does not have, according to Arnett?s studies, correlations with any other measurements.

La prolifération des capteurs d'images dans de nombreux appareils électroniques, et l'évolution des capacités de traitements à proximité de ces capteurs ouvrent un champ d'exploration pour l'implantation et l'optimisation d'algorithmes complexes de traitement d'images afin de proposer des systèmes de vision artificielle embarquée. Ces travaux s'inscrivent dans la problématique dite d'adéquation algorithme-architecture (A3). Ils portent sur une classe d'algorithmes appelée réseau de neurones à convolutions (ConvNet) et ses applications en analyse de visages embarquée. La chaîne d'analyse de visages, introduite par Garcia et al., a été choisie d'une part pour ses performances en taux de détection/reconnaissance au niveau de l'état de l'art, et d'autre part pour son caractère homogène reposant sur des ConvNets. La première contribution de ces travaux porte sur une étude d'adéquation de cette chaîne d'analyse de visages aux processeurs embarqués. Nous proposons plusieurs adaptations algorithmiques des ConvNets, et montrons que celles-ci permettent d'obtenir des facteurs d'accélération importants (jusqu'à 700) sur un processeur embarqué pour mobile, sans dégradation des performances en taux de détection/reconnaissance. Nous présentons ensuite une étude des capacités de parallélisation des ConvNets, au travers des travaux de thèse de N. Farrugia. Une exploration "gros-grain" du parallélisme des ConvNets, suivie d'une étude de l'ordonnancement interne des processeurs élémentaires, conduisent à une architecture parallèle paramétrable, capable de détecter des visages à plus de 10 images VGA par seconde sur FPGA. Nous proposons enfin une extension de ces études à la phase d'apprentissage de ces réseaux de neurones. Nous étudions des restrictions de l'espace des hypothèses d'apprentissage, et montrons, sur un cas d'application, que les capacités d'apprentissage des ConvNets ne sont pas dégradées, et que le temps d'apprentissage peut être réduit jusqu'à un facteur cinq. / Proliferation of image sensors in many electronic devices, and increasing processing capabilities of such sensors, open a field of exploration for the implementation and optimization of complex image processing algorithms in order to provide embedded vision systems. This work is a contribution in the research domain of algorithm-architecture matching. It focuses on a class of algorithms called convolution neural network (ConvNet) and its applications in embedded facial analysis. The facial analysis framework, introduced by Garcia et al., was chosen for its state of the art performances in detection/recognition, and also for its homogeneity based on ConvNets. The first contribution of this work deals with an adequacy study of this facial analysis framework with embedded processors. We propose several algorithmic adaptations of ConvNets, and show that they can lead to significant speedup factors (up to 700) on an embedded processor for mobile phone, without performance degradation. We then present a study of ConvNets parallelization capabilities, through N. Farrugia's PhD work. A coarse-grain parallelism exploration of ConvNets, followed by study of internal scheduling of elementary processors, lead to a parameterized parallel architecture on FPGA, able to detect faces at more than 10 VGA frames per second. Finally, we propose an extension of these studies to the learning phase of neural networks. We analyze several hypothesis space restrictions for ConvNets, and show, on a case study, that classification rate performances are almost the same with a training time divided by up to five.

Dans cette thèse, nous explorons l'utilisation des dérivées Gaussiennes multi-échelles comme représentation initiale pour la détection, la reconnaissance et la classification des visages humains dans des images. Nous montrons qu'un algorithme rapide, $O(N)$, de construction d'une pyramide binomiale peut être utilisé pour extraire des dérivées Gaussiennes avec une réponse impulsionnelle identique à un facteur d'échelle $sqrt{2}$>. Nous montrons ensuite qu'un vecteur composé de ces dérivées à différentes échelles et à différents ordres en chaque pixel peut être utilisé comme base pour les algorithmes de détection, de classification et de reconnaissance lesquels atteignent ou dépassent les performances de l'état de l'art avec un coût de calcul réduit. De plus l'utilisation de coefficients entiers, avec une complexité de calcul et des exigences mémoires en $O(N)$ font qu'une telle approche est appropriée pour des applications temps réel embarquées sur des systèmes mobiles. Nous testons cette représentation en utilisant trois problèmes classiques d'analyse d'images faciales : détection de visages, reconnaissance de visages et estimation de l'âge. Pour la détection de visages, nous examinons les dérivées Gaussiennes multi-échelles comme une alternative aux ondelettes de Haar pour une utilisation dans la construction d'une cascade de classifieurs linéaires appris avec l'algorithme Adaboost, popularisé par Viola and Jones. Nous montrons que la représentation pyramidale peut être utilisée pour optimiser le processus de détection en adaptant la position des dérivées dans la cascade. Dans ces experiences nous sommes capables de montrer que nous pouvons obtenir des niveaux de performances de détection similaires (mesurés par des courbes ROC) avec une réduction importante du coût de calcul. Pour la reconnaissance de visages et l'estimation de l'âge, nous montrons que les dérivées Gaussiennes multi-échelles peuvent être utilisées pour calculer une représentation tensorielle qui conserve l'information faciale la plus importante. Nous montrons que combinée à l'Analyse Multilinéaire en Composantes Principales et à la méthode Kernel Discriminative Common Vectors (KDCV), cette représentation tensorielle peut mener à un algorithme qui est similaire aux techniques concurrentes pour la reconnaissance de visages avec un coût de calcul réduit. Pour l'estimation de l'âge à partir d'images faciales, nous montrons que notre représentation tensorielle utilisant les dérivées de Gaussiennes multi-échelles peut être utilisée avec une machine à vecteur de pertinence pour fournir une estimation de l'âge avec des niveaux de performances similaires aux méthodes de l'état de l'art. / In this thesis, we propose to modelize facial images using Gaussian Derivatives computed with a Half-Octave Gaussian Pyramid. In this scope, Gaussian derivatives have shown a high versatility in object recognition and image analysis, nevertheless there is not a considerable number of proposed aproaches in the state-of-the-art that uses Gaussian derivatives for extracting important information from facial images. Motivated by the above mentioned and the high amount of applications in facial analysis, security systems and Biometry, in this thesis as a first time, we propose to use an unique image representation, the Gaussian Scale Space computed with a half octave pyramid. We show in this thesis that this image representation could be used to perform different tasks in facial analysis without lost of performance compared with other approaches in the state-of-the-art that uses more complicated image representations. it is also well know that using an unique image represenation could be convenient in real world applications where the amount of memory capacity is limitated by hardware constraints. To demostrate our assumptations we solve three different tasks in facial analysis: Face detection, Face recognition and Age estimation. In face detection we propose to use a cascade of classifiers using Gaussian derivatives. Specifically we propose to use Gaussian derivatives up to the fourth order, in effect experiemnts using different derivatives orders have shown that fourth order Gaussian derivatives provide important information in face detection and recognition. In adition, to improve the speed of detection using Gaussian derivatives, we develope a new cascade architecture which considerates the computational cost of each Gaussian derivative order to chose its best position in the cascade. Finally, to solve the face recognition and age estimation problems, we propose a tensorial model based in Gaussian derivatives. This tensorial model preserves the 3-D structure of feature space and it does not break the natural structure of data when a vectorization process is applied. Each one of the methods proposed in the thesis are discused and validated with a set of well defined experiments. All our results are compared with the last state-of-the-art results in face detection, recognition and age estimation, giving comparable or superior results