Recently, ocular biometrics in unconstrained environments using images obtained at visible wavelength have gained the researchers' attention, especially with images captured by mobile devices. Periocular recognition has been demonstrated to be an alternative when the iris trait is not available due to occlusions or low image resolution. However, the periocular trait does not have the high uniqueness presented in the iris trait. Thus, the use of datasets containing many subjects is essential to assess biometric systems' capacity to extract discriminating information from the periocular region. Also, to address the within-class variability caused by lighting and attributes in the periocular region, it is of paramount importance to use datasets with images of the same subject captured in distinct sessions. As the datasets available in the literature do not present all these factors, in this work, we present a new periocular dataset containing samples from 1,122 subjects, acquired in 3 sessions by 196 different mobile devices. The images were captured under unconstrained environments with just a single instruction to the participants: to place their eyes on a region of interest. We also performed an extensive benchmark with several Convolutional Neural Network (CNN) architectures and models that have been employed in state-ofthe-art approaches based on Multi-class Classification, Multitask Learning, Pairwise Filters Network, and Siamese Network. The results achieved in the closed-and open-world protocol, considering the identification and verification tasks, show that this area still needs research and development.

Tese apresentada como parte dos requisitos para obtenção do Grau de Doutor em Ciências na Área de Tecnologia Nuclear -Aplicações. utilizando a espectrometria como padrão de referência, com incertezas relativas de 0,62% para k=1. As incertezas associadas ao método proposto foram satisfatórias, para um laboratório padrão secundário e comparáveis a um laboratório primário. were found with differences from that provided by the primary laboratory between 1.3% and 15.31%. The results indicated the feasibility of establishing a protocol for the calibration of instruments using spectrometry as a reference standard, with relative uncertainties of 0.62% for k = 1. The uncertainties associated with the proposed method were satisfactory for a secondary standard laboratory and comparable to a primary laboratory. Orientadora: Profª. Drª. Maria da Penha Albuquerque Potiens SÃO PAULO 2017 INSTITUTO DE PESQUISAS ENERGÉTICAS E NUCLEARES Autarquia associada à Universidade de São Paulo DESENVOLVIMENTO DE UM PROTOCOLO DE CALIBRAÇÃO UTILIZANDO ESPECTROMETRIA E SIMULAÇÃO MATEMÁTICA, EM FEIXES PADRÕES DE RAIOS X LUCAS RODIRGUES DOS SANTOS DEVELOPMENT OF A CALIBRATION PROTOCOL USING SPECTROMETRY AND MATHEMATICAL SIMULATION, IN X RAY STANDARD BEAMS

No abstract

Introduction: The discovery of ionizing radiation and its application in medicine was an important advance for society due to the various possibilities of use for treatment and diagnosis of diseases. On the other hand, the knowledge about the damages caused by the biological effects of the ionizing radiation, requires a continuous improvement of the quality control in diagnostic radiology. Radiation detectors are equipment used to measure the levels of radiation emitted by sources, be they natural or artificial. Among the most commonly used detectors are ionization chambers (IC). During transport for clinical routine use, sudden changes in temperature and humidity may affect the behavior of these detectors at the time of measurement. For the equipment to recompose its stability, it is necessary a time that is often incompatible with the free time available for the tests. The objective of this work was to evaluate the performance of an ionization chamber used in radioprotection measurements in diagnostic radiology, considering the climatic variations in the different regions of Brazil[1]. For that, a system of air conditioning was developed to simulate the environmental situations found for the parameters of temperature and humidity at the moment of the clinical measurements, making possible to estimate the influence of these factors in the obtained values.Methods: Before submitting the IC to the proposed climatic tests, the dosimetric system composed of an ionization chamber for radioprotection (Radcal®, 10x15 - 1800) and electrometer (Radcal® 9015) was calibrated in the instrument calibration laboratory of the Instituto de Pesquisas Energéticas e Nucleares (LCI-IPEN), in the ISO N series radiation qualities[2,3]. In order to evaluate the behavior of the detector in different environmental conditions, a system of air conditioning was constructed, which allows the control of humidity and temperature in an acrylic box with dimensions appropriate to the test, and a control source of 90Sr + 90Y [PTW® , model 8921, 33MBq (1994)] for the irradiation of IC. The IC was then positioned within the acclimatization chamber for 15 minutes for stabilization according to the manufacturer's manual. A series of measurements was made with the source of the check to analyze the behavior of the CI with the typical environmental conditions of the city of Manaus, capital of Amazonas state that is located in the north of Brazil.Conclusions: Preliminary results indicate an increase in the accumulated dose values obtained with the ionization chamber due to increase in temperature. The humidity parameter interfered in the background value measured by the ionization chamber and after the test, the ionization chamber showed a significant current leakage. Determining the influence of environmental factors is not an easy task, but it is fundamental to the reliability of measurement and optimization of quality control. The methodology proposed in this work pointed to a repeatable character, allowing the simulation of environmental tests for the parameters of humidity and temperature of several models of ionization chambers.

EPÍGRAFE "Conhecer e pensar não é chegar a uma verdade absolutamente certa, mas dialogar com a incerteza" (Edgar Morin, 2000) MONITORAÇÂO IN VIVO-ANÁLISE DE INCERTEZAS Lucas Rodrigues dos Santos RESUMO Este trabalho teve como objetivos estabelecer os protocolos para o cálculo de incertezas de medição e a interpretação dos resultados de monitoração do Laboratório de Monitoração In Vivo (LMIV). Especificamente, o trabalho visou quantificar a contribuição dos fatores de influência para o cálculo de incertezas, e estabelecer a incerteza nas medidas. Foram utilizados dois detectores cintiladores de iodeto de sódio ativados com tálio (NaI:Tl) que possuem dimensões de 8x4" (detector de corpo inteiro) e 3x3" (detector de tireóide), um analisador multicanal EG&G Ortec, modelo 920E, e um microcomputador onde os espectros são adquiridos, analisados e armazenados com o auxílio do programa Renascence32, Ortec. As medições foram realizadas utilizando-se o objeto simulador antropomórfico do Alderson Research Labs. O radionuclídeo de interesse adotado para este estudo foi o 137 Cs. Foram estudadas as influências dos fatores operador, geometria de medição, condições ambientais, flutuações eletrônicas com o tempo e baixa atividade. A análise das incertezas resultou em uma incerteza combinada relativa de 15,7% para o sistema 8x4" e 9,8% para o sistema 3x3". Estes valores foram obtidos seguindo os princípios recomendados no Guia para Expressão da Incerteza de Medição (GUM) da Joint Committee for Guides in Metrology (JCGM).