Estimación eficiente de atributos demográficos del rostro humano en imágenes

Sin duda, el rostro humano ofrece mucha más información de la que pensamos. La cara transmite sin nuestro consentimiento señales no verbales, a partir de las interacciones faciales, que dejan al descubierto nuestro estado afectivo, actividad cognitiva, personalidad y enfermedades. Estudios recientes [OFT14, TODMS15] demuestran que muchas de nuestras decisiones sociales e interpersonales derivan de un previo análisis facial de la cara que nos permite establecer si esa persona es confiable, trabajadora, inteligente, etc. Esta interpretación, propensa a errores, deriva de la capacidad innata de los seres humanas de encontrar estas señales e interpretarlas. Esta capacidad es motivo de estudio, con un especial interés en desarrollar métodos que tengan la habilidad de calcular de manera automática estas señales o atributos asociados a la cara. Así, el interés por la estimación de atributos faciales ha crecido rápidamente en los últimos años por las diversas aplicaciones en que estos métodos pueden ser utilizados: marketing dirigido, sistemas de seguridad, interacción hombre-máquina, etc. Sin embargo, éstos están lejos de ser perfectos y robustos en cualquier dominio de problemas. La principal dificultad encontrada es causada por la alta variabilidad intra-clase debida a los cambios en la condición de la imagen: cambios de iluminación, oclusiones, expresiones faciales, edad, género, etnia, etc.; encontradas frecuentemente en imágenes adquiridas en entornos no controlados. Este de trabajo de investigación estudia técnicas de análisis de imágenes para estimar atributos faciales como el género, la edad y la postura, empleando métodos lineales y explotando las dependencias estadísticas entre estos atributos. Adicionalmente, nuestra propuesta se centrará en la construcción de estimadores que tengan una fuerte relación entre rendimiento y coste computacional. Con respecto a éste último punto, estudiamos un conjunto de estrategias para la clasificación de género y las comparamos con una propuesta basada en un clasificador Bayesiano y una adecuada extracción de características. Analizamos en profundidad el motivo de porqué las técnicas lineales no han logrado resultados competitivos hasta la fecha y mostramos cómo obtener rendimientos similares a las mejores técnicas no-lineales. Se propone un segundo algoritmo para la estimación de edad, basado en un regresor K-NN y una adecuada selección de características tal como se propuso para la clasificación de género. A partir de los experimentos desarrollados, observamos que el rendimiento de los clasificadores se reduce significativamente si los ´estos han sido entrenados y probados sobre diferentes bases de datos. Hemos encontrado que una de las causas es la existencia de dependencias entre atributos faciales que no han sido consideradas en la construcción de los clasificadores. Nuestro resultados demuestran que la variabilidad intra-clase puede ser reducida cuando se consideran las dependencias estadísticas entre los atributos faciales de el género, la edad y la pose; mejorando el rendimiento de nuestros clasificadores de atributos faciales con un coste computacional pequeño. Abstract Surely the human face provides much more information than we think. The face provides without our consent nonverbal cues from facial interactions that reveal our emotional state, cognitive activity, personality and disease. Recent studies [OFT14, TODMS15] show that many of our social and interpersonal decisions derive from a previous facial analysis that allows us to establish whether that person is trustworthy, hardworking, intelligent, etc. This error-prone interpretation derives from the innate ability of human beings to find and interpret these signals. This capability is being studied, with a special interest in developing methods that have the ability to automatically calculate these signs or attributes associated with the face. Thus, the interest in the estimation of facial attributes has grown rapidly in recent years by the various applications in which these methods can be used: targeted marketing, security systems, human-computer interaction, etc. However, these are far from being perfect and robust in any domain of problems. The main difficulty encountered is caused by the high intra-class variability due to changes in the condition of the image: lighting changes, occlusions, facial expressions, age, gender, ethnicity, etc.; often found in images acquired in uncontrolled environments. This research work studies image analysis techniques to estimate facial attributes such as gender, age and pose, using linear methods, and exploiting the statistical dependencies between these attributes. In addition, our proposal will focus on the construction of classifiers that have a good balance between performance and computational cost. We studied a set of strategies for gender classification and we compare them with a proposal based on a Bayesian classifier and a suitable feature extraction based on Linear Discriminant Analysis. We study in depth why linear techniques have failed to provide competitive results to date and show how to obtain similar performances to the best non-linear techniques. A second algorithm is proposed for estimating age, which is based on a K-NN regressor and proper selection of features such as those proposed for the classification of gender. From our experiments we note that performance estimates are significantly reduced if they have been trained and tested on different databases. We have found that one of the causes is the existence of dependencies between facial features that have not been considered in the construction of classifiers. Our results demonstrate that intra-class variability can be reduced when considering the statistical dependencies between facial attributes gender, age and pose, thus improving the performance of our classifiers with a reduced computational cost.

Detection of Biological CO2 and 1,3-Pentadiene Using Non-refrigerated Low-Cost MWIR Detectors

The early detection of spoiling metabolic products in contaminated food is a very important tool to control quality. Some volatile compounds produce unpleasant odours at very low concentrations, making their early detection very challenging. This is the case of 1,3-pentadiene produced by microorganisms through decarboxylation of the preservative sorbate. In this work, we have developed a methodology to use the data produced by a low-cost, compact MWIR (Mid-Wave IR) spectrometry device without moving parts, which is based on a linear array of 128 elements of VPD PbSe coupled to a linear variable filter (LVF) working in the spectral range between 3 and 4.6 ?m. This device is able to analyze food headspace gases through dedicated sample presentation setup. This methodology enables the detection of CO2 and the volatile compound 1,3-pentadiene, as compared to synthetic patrons. Data analysis is based on an automated multidimensional dynamic processing of the MWIR spectra. Principal component and discriminant analysis allow segregating between four yeast strains including producers and no producers. The segregation power is accounted as a measure of the discrimination quality.

Geoquímica multielementar de sedimentos de corrente no Estado de São Paulo: abordagem através da análise estatística multivariada

Este trabalho apresenta resultados geoquímicos multielementares de sedimentos de corrente no estado de São Paulo, obtidos através do projeto institucional do Serviço Geológico do Brasil denominado \"Levantamento Geoquímico de Baixa Densidade no Brasil\". Dados analíticos de 1422 amostras de sedimento de corrente obtidos por ICP-MS (Inductively Coupled Plasma Mass Spectrometry), para 32 elementos químicos (Al, Ba, Be, Ca, Ce, Co, Cr, Cs, Cu, Fe, Ga, Hf, K, La, Mg, Mn, Mo, Nb, Ni, P, Pb, Rb, Sc, Sn, Sr, Th, Ti, U, V, Y, Zn e Zr), foram processadas e abordadas através da análise estatística uni e multivariada. Os resultados do tratamento dos dados através de técnicas estatísticas univariadas forneceram os valores de background geoquímico (teor de fundo) dos 32 elementos para todo estado de São Paulo. A análise georreferenciada das distribuições geoquímicas unielementares evidenciaram a compartimentação geológica da área. As duas principais províncias geológicas do estado de São Paulo, Bacia do Paraná e Complexo Cristalino, se destacam claramente na maioria das distribuições geoquímicas. Unidades geológicas de maior expressão, como a Formação Serra Geral e o Grupo Bauru também foram claramente destacadas. Outras feições geoquímicas indicaram possíveis áreas contaminadas e unidades geológicas não cartografadas. Os resultados da aplicação de métodos estatísticos multivariados aos dados geoquímicos com 24 variáveis (Al, Ba, Ce, Co, Cr, Cs, Cu, Fe, Ga, La, Mn, Nb, Ni, Pb, Rb, Sc, Sr, Th, Ti, U, V, Y, Zn e Zr) permitiram definir as principais assinaturas e associações geoquímicas existentes em todo estado de São Paulo e correlacioná-las aos principais domínios litológicos. A análise de agrupamentos em modo Q forneceu oito grupos de amostras geoquimicamente correlacionáveis, que georreferenciadas reproduziram os principais compartimentos geológicos do estado: Complexo Cristalino, Grupos Itararé e Passa Dois, Formação Serra Geral e Grupos Bauru e Caiuá. A análise discriminante multigrupos comprovou, estatisticamente, a classificação dos grupos formados pela análise de agrupamentos e forneceu as principais variáveis discriminantes: Fe, Co, Sc, V e Cu. A análise de componentes principais, abordada em conjunto com a análise fatorial pelo método de rotação varimax, forneceram os principais fatores multivariados e suas respectivas associações elementares. O georreferenciamento dos valores de escores fatoriais multivariados delimitaram as áreas onde as associações elementares ocorrem e forneceram mapas multivariados para todo o estado. Por fim, conclui-se que os métodos estatísticos aplicados são indispensáveis no tratamento, apresentação e interpretação de dados geoquímicos. Ademais, com base em uma visão integrada dos resultados obtidos, este trabalho recomenda: (1) a execução dos levantamentos geoquímicos de baixa densidade em todo país em caráter de prioridade, pois são altamente eficazes na definição de backgrounds regionais e delimitação de províncias geoquímicas com interesse metalogenético e ambiental; (2) a execução do mapeamento geológico contínuo em escala adequada (maiores que 1:100.000) em áreas que apontam para possíveis existências de unidades não cartografadas nos mapas geológicos atuais.

Variables sociodemográficas, médicas y psicológicas relacionadas con el comportamiento de fumar y con los intentos de dejar de fumar en una muestra comunitaria

Se estudia la relación entre variables sociodemográficas, médicas y psicológicas y el comportamiento de fumar y los intentos de dejar de fumar. La muestra (n=291) se ha extraído al azar de una población comunitaria rural de 4300 habitantes. Todas las variables se han medido con un único cuestionario, a través de entrevista personal domiciliaria. Mediante análisis discriminantes las variables sexo, consumo de bebidas alcohólicas y el uso de medicamentos son las que más explican estadísticamente el comportamiento de fumar. Sorprendentemente, sólo una variable, haber visitado al médico en los últimos 12 meses, se asoció, bivariadamente, con los intentos de dejar de fumar, y no se procedía, lógicamente, con el análisis multivariante. Por último, se discuten los hallazgos a la luz de la literatura internacional y nacional.

Morphological classification of microtidal sand and gravel beaches

A new classification of microtidal sand and gravel beaches with very different morphologies is presented below. In 557 studied transects, 14 variables were used. Among the variables to be emphasized is the depth of the Posidonia oceanica. The classification was performed for 9 types of beaches: Type 1: Sand and gravel beaches, Type 2: Sand and gravel separated beaches, Type 3: Gravel and sand beaches, Type 4: Gravel and sand separated beaches, Type 5: Pure gravel beaches, Type 6: Open sand beaches, Type 7: Supported sand beaches, Type 8: Bisupported sand beaches and Type 9: Enclosed beaches. For the classification, several tools were used: discriminant analysis, neural networks and Support Vector Machines (SVM), the results were then compared. As there is no theory for deciding which is the most convenient neural network architecture to deal with a particular data set, an experimental study was performed with different numbers of neuron in the hidden layer. Finally, an architecture with 30 neurons was chosen. Different kernels were employed for SVM (Linear, Polynomial, Radial basis function and Sigmoid). The results obtained for the discriminant analysis were not as good as those obtained for the other two methods (ANN and SVM) which showed similar success.

The SAAS-R: A New Instrument to Assess the School Attitudes of Students with High and Low Academic Achievement in Spain

We examined the psychometric properties of the School Attitude Assessment Survey–Revised in a Spanish population (n = 1,398). Confirmatory factor analysis procedures supported the instrument’s five-factor structure. The results of discriminant analysis demonstrated the predictive power of the School Attitude Assessment Survey–Revised scales as regards academic performance. Implications for education and assessment are discussed.

(Table 1) Summary statistics of taxa in samples of ODP Hole 181-1123B

Planktonic foraminifera are used to identify late Pliocene-Quaternary near surface water masses on the northeastern flank of Chatham Rise by comparison with faunas in core-tops east of New Zealand. In an overview study, distance measures, ordinations, and discriminant analysis are applied to 32 faunas from Site 1123B to identify similar faunas among 35 core-tops between 35 and 61°S east of New Zealand. Many Site 1123B faunas in the 2.72 myr interval sampled compare with those in core-tops on the northern side of Chatham Rise from a similar latitude, and are identified as transitional zone assemblages now associated with the subtropical gyre. This result is consistent with studies of late Quaternary planktonic foraminifera from this region and suggests that, typically, the Subtropical Front was locked to Chatham Rise through glacial and interglacial periods, at least back to the late Pliocene. However, a fauna at ca. 1.17 Ma compares with subpolar assemblages in core-tops between 44 and 48°S and identifies cooler surface water. Expectedly, closer sampling may reveal additional periods when southern water moved over the northeastern flank of Chatham Rise. Although the dominance of Globorotalia inflata, a species typical of the southern margin of subtropical gyres, is a principal feature of Site 1123B faunas, in a minority it is replaced as the most abundant species by dextral populations of Neogloboquadrina pachyderma, particularly about the time of the middle Pleistocene transition. Close analogues of these variant transitional assemblages are not present in core-tops about Chatham Rise but sediment trap and coretop data from other regions suggest that they identify high fertility in the mixed layer associated with upwelling or mixing of water masses. The proportion of sinistrally coiled Neogloboquadrina pachyderma rises to ca. 0.6 between ca. 2.45 and 2.57 Ma, soon after the intensification of Northern Hemisphere glaciation. Although the coiling data indicate subantarctic near surface water, the species remains rare. As the faunas retain their transitional zone character, only minor entrainment of subantarctic water may have occurred.

Optimized Decoding for Auditory Attention Detection

Thesis (Master's)--University of Washington, 2016-06

Computational properties of min/max autocorrelation factors

Minimum/maximum autocorrelation factor (MAF) is a suitable algorithm for orthogonalization of a vector random field. Orthogonalization avoids the use of multivariate geostatistics during joint stochastic modeling of geological attributes. This manuscript demonstrates in a practical way that computation of MAF is the same as discriminant analysis of the nested structures. Mathematica software is used to illustrate MAF calculations from a linear model of coregionalization (LMC) model. The limitation of two nested structures in the LMC for MAF is also discussed and linked to the effects of anisotropy and support. The analysis elucidates the matrix properties behind the approach and clarifies relationships that may be useful for model-based approaches. (C) 2003 Elsevier Science Ltd. All rights reserved.

Investigation of psychological and neuropsychological factors associated with clinical outcome following a group rehabilitation programme

The present study investigated neuropsychological and psychological factors associated with successful treatment outcome following a group intervention for individuals with acquired brain injury (ABI). Participants were classified into two groups (Clinically Improved and Not Improved) based upon the findings of a previous study (Ownsworth, McFarland, & Young, 2000a). A discriminant analysis was used to predict group membership on three outcome measures (Awareness and Strategy Behaviour indices of the Self-Regulation Skills Interview and the Psychosocial Dimension of the Sickness Impact Profile) between pre-assessment and post-assessment, and between pre-assessment and 6 months follow-up. Neuropsychological factors involved measures of executive functioning and psychological factors were assessed using measures of personality-related denial and coping-related denial. Overall, the results indicated that individuals with impaired executive functioning were most likely to be classified as Clinically Improved on measures of awareness, strategy behaviour and psychosocial functioning. Individuals who deny or minimise their ABI symptoms were less likely to improve their psychosocial functioning following the group intervention. Future research needs to evaluate interventions for enhancing self-regulation skills and improving psychosocial functioning for individuals who employ denial as a main strategy for coping following ABI.

Overlapping species boundaries and hybridization within the Monastraea "annularis" reef coral complex in the Pleistocene of the Bahama Islands

Recent molecular analyses indicate that many reef coral species belong to hybridizing species complexes or "syngameons." Such complexes consist of numerous genetically distinct-species or lineages, which periodically split and/or fuse as they extend through time. During splitting and fusion, morphologic intermediates form and species overlap. Here we focus on processes associated with lineage fusion, specifically introgressive hybridization, and the recognition of such hybridization in the fossil record. Our approach involves comparing patterns of ecologic and morphologic overlap in genetically characterized modern species with fossil representatives of the same or closely related species. We similarly consider the long-term consequences of past hybridization on the structure of modern-day species boundaries. Our study involves the species complex Montastraea annularis s.l. and is based in the Bahamas, where, unlike other Caribbean locations, two of the three members of the complex today are not genetically distinct. We measured and collected colonies along linear transects across Pleistocene reef terraces of last interglacial age (approximately 125 Ka) on the islands of San Salvador, Andros, and Great Inagua. We performed quantitative ecologic and morphologic analyses of the fossil data, and compared patterns of overlap among species with data from modern localities where species are and are not genetically distinct. Ecologic and morphologic analyses reveal "moderate" overlap (>10%, but statistically significant differences) and sometimes "high" overlap (no statistically significant differences) among Pleistocene growth forms (= "species"). Ecologic analyses show that three species (massive, column, organ-pipe) co-occurred. Although organ-pipes had higher abundances in patch reef environments, columnar and massive species exhibited broad, completely overlapping distributions and had abundances that were not related to reef environment. For morphometric analyses, we used multivariate discriminant analysis on landmark data and linear measurements. The results show that columnar species overlap "moderately" with organ-pipe and massive species. Comparisons with genetically characterized colonies from Panama show that the Pleistocene Bahamas species have intermediate morphologies, and that the observed "moderate" overlap differs from the morphologic separation among the three modern species. In contrast, massive and columnar species from the Pleistocene of the Dominican Republic comprise distinct morphologic clusters, similar to the modern species; organ-pipe species exhibit "low" overlap (

Clinical tests of musculoskeletal dysfunction in the diagnosis of cervicogenic headache

Persistent intermittent headache is a common disorder and is often accompanied by neck aching or stiffness, which could infer a cervical contribution to headache. However, the incidence of cervicogenic headache is estimated to be 14-18% of all chronic headaches, highlighting the need for clear criterion of cervical musculoskeletal impairment to identify cervicogenic headache sufferers who may benefit from treatments such as manual therapy. This study examined the presence of cervical musculoskeletal impairment in 77 subjects, 27 with cervicogenic headache, 25 with migraine with aura and 25 control subjects. Assessments included a photographic measure of posture, range of movement, cervical manual examination, pressure pain thresholds, muscle length, performance in the cranio-cervical flexion test and cervical kinaesthetic sense. The results indicated that when compared to the migraine with aura and control groups who scored similarly in the tests, the cervicogenic headache group had less range of cervical flexion/extension (P = 0.048) and significantly higher incidences of painful upper cervical joint dysfunction assessed by manual examination (all P < 0.05) and muscle tightness (P < 0.05). Sternocleidomastoid normalized EMG values were higher in the latter three stages of the cranio-cervical flexion test although they failed to reach significance. There were no between group differences for other measures. A discriminant analysis revealed that manual examination could discriminate the cervicogenic headache group from the other subjects (migraine with aura and control subjects combined) with an 80% sensitivity. (C) 2005 Elsevier Ltd. All rights reserved.

Improvement in measures of psychological distress amongst amphetamine misusers treated with brief cognitive-behavioural therapy (CBT)

This trial of cognitive-behavioural therapy (CBT) based amphetamine abstinence program (n = 507) focused on refusal self-efficacy, improved coping, improved problem solving and planning for relapse prevention. Measures included the Severity of Dependence Scale (SDS), the General Health Questionnaire-28 (GHQ-28) and Amphetamine Refusal Self-Efficacy. Psychiatric case identification (caseness) across the four GHQ-28 sub-scales was compared with Australian normative data. Almost 90% were amphetamine-dependent (SDS 8.15 +/- 3.17). Pretreatment, all GHQ-28 sub-scale measures were below reported Australian population values. Caseness was substantially higher than Australian normative values {Somatic Symptoms (52.3%), Anxiety (68%), Social Dysfunction (46.5%) and Depression (33.7%). One hundred and sixty-eight subjects (33%) completed and reported program abstinence. Program completers reported improvement across all GHQ-28 sub-scales Somatic Symptoms (p < 0.001), Anxiety (p < 0.001), Social Dysfunction (p < 0.001) and Depression (p < 0.001)}. They also reported improvement in amphetamine refusal self-efficacy (p < 0.001). Improvement remained significant following intention-to-treat analyses, imputing baseline data for subjects that withdrew from the program. The GHQ-28 sub-scales, Amphetamine Refusal Self-Efficacy Questionnaire and the SDS successfully predicted treatment compliance through a discriminant analysis function (p

Comparison of ASAS improvement criteria for ankylosing spondylitis with clinically relevant improvement according to an expert panel

Objective: To investigate whether the recently developed (statistically derived) "ASsessment in Ankylosing Spondylitis Working Group" improvement criteria (ASAS-IC) for ankylosing spondylitis (AS) reflect clinically relevant improvement according to the opinion of an expert panel. Methods: The ASAS-IC consist of four domains: physical function, spinal pain, patient global assessment, and inflammation. Scores on these four domains of 55 patients with AS, who had participated in a non-steroidal anti-inflammatory drug efficacy trial, were presented to an international expert panel (consisting of patients with AS and members of the ASAS Working Group) in a three round Delphi exercise. The number of (non-) responders according to the ASAS-IC was compared with the final-consensus of the experts. The most important domains in the opinion of the experts were identified, and also selected with discriminant analysis. A number of provisional criteria sets that best represented the consensus of the experts were defined. Using other datasets, these clinically derived criteria sets as well as the statistically derived ASAS-IC were then tested for discriminative properties and for agreement with the end of trial efficacy by patient and doctor. Results: Forty experts completed the three Delphi rounds. The experts considered twice as many patients to be responders than the ASAS-IC (42 v 21). Overall agreement between experts and ASAS-IC was 62%. Spinal pain was considered the most important domain by most experts and was also selected as such by discriminant analysis. Provisional criteria sets with an agreement of greater than or equal to 80% compared with the consensus of the experts showed high placebo response rates (27-42%), in contrast with the ASAS-IC with a predefined placebo response rate of 25%. All criteria sets and the ASAS-IC discriminated well between active and placebo treatment (chi(2) = 36-45; p < 0.001). Compared with the end of trial efficacy assessment, the provisional criteria sets showed an agreement of 71-82%, sensitivity of 67-83%, and specificity of 81-88%. The ASAS-IC showed an agreement of 70%, sensitivity of 62%, and specificity of 89%. Conclusion: The ASAS-IC are strict in defining response, are highly specific, and consequently show lower sensitivity than the clinically derived criteria sets. However, those patients who are considered as responders by applying the ASAS-IC are acknowledged as such by the expert panel as well as by. patients' and doctors' judgments, and are therefore likely to be true responders.

Exploratory database visualisation: the application and assessment of data and dimensionality reduction

This thesis describes the development of a complete data visualisation system for large tabular databases, such as those commonly found in a business environment. A state-of-the-art 'cyberspace cell' data visualisation technique was investigated and a powerful visualisation system using it was implemented. Although allowing databases to be explored and conclusions drawn, it had several drawbacks, the majority of which were due to the three-dimensional nature of the visualisation. A novel two-dimensional generic visualisation system, known as MADEN, was then developed and implemented, based upon a 2-D matrix of 'density plots'. MADEN allows an entire high-dimensional database to be visualised in one window, while permitting close analysis in 'enlargement' windows. Selections of records can be made and examined, and dependencies between fields can be investigated in detail. MADEN was used as a tool for investigating and assessing many data processing algorithms, firstly data-reducing (clustering) methods, then dimensionality-reducing techniques. These included a new 'directed' form of principal components analysis, several novel applications of artificial neural networks, and discriminant analysis techniques which illustrated how groups within a database can be separated. To illustrate the power of the system, MADEN was used to explore customer databases from two financial institutions, resulting in a number of discoveries which would be of interest to a marketing manager. Finally, the database of results from the 1992 UK Research Assessment Exercise was analysed. Using MADEN allowed both universities and disciplines to be graphically compared, and supplied some startling revelations, including empirical evidence of the 'Oxbridge factor'.