Malgré sa capacité unique à discerner des détails qu'aucun instrument “classique” ne peut voir, l'interférométrie optique est fortement handicapée par l'atmosphère. Celle-ci limite drastiquement les temps de pose des interféromètres au sol et les empèche d'accumuler suffisamment de photons pour observer des sources toujours plus faibles, limitant de facto l'échantillon des astres observables. Les suiveurs de franges sont des instruments développés spécifiquement dans le but de compenser ces perturbations atmosphériques, et ainsi de repousser les limites de l'univers visible par les interféromètres optiques. Le but premier de cette thèse est d'étudier et d'améliorer ces instruments, dans le contexte des technologies actuelles et des nouvelles générations d'interféromètres combinant 4 télescopes et plus. La seconde grande partie de cette thèse s'attachera quant à elle à montrer l'intérêt de l'interférométrie optique dans l'étude des binaires en interaction, astres en mesure d'apporter des réponses à un vaste panel de champs d'étude du fait de la diversité des processus physiques en jeu en leur sein. / Despite its unique ability to discern details that a “classical” instrument cannot see, optical interferometry is still strongly handicapped by the atmosphere. It drastically limits the exposure time of ground interferometers and prevents them to accumulate enough photons to observe weak sources, limiting de facto sample of observable stars. Fringe trackers are instruments developed specifically to compensate for these atmospheric disturbances, and so push the boundaries of the universe observable with optical interferometers. The primary purpose of this thesis is to study and improve these instruments in the context of the current technologies and of the new generation of interferometers combining four telescopes and more. The second major part of this thesis will show the advantages of optical interferometry in the study of interacting binary, stars able to answer to a wide range of domains because of the diversity of physical processes involved in them.

The assumption in economics of a representative agent is often made. However, it is a very rigid assumption. Hall and Jones (2004b) presented an economic model that essentially provided for a representative agent for each age group in determining the group's health level function. Our work seeks to extend their theoretical version of the model by allowing for two representative agents for each age—one for each of “Healthy” and “Sick” risk-factor groups—to allow for additional heterogeneity in the populace. The approach to include even more risk-factor groups is also briefly discussed. While our “extended” theoretical model is not applied directly to relevant data, several techniques that could be applicable were the relevant data to be obtained are demonstrated on other data sets. This includes examples of using linear classification, fitting baseline-category logit models, and running the genetic algorithm.

This study examined the efficacy of the Sound Foundations program, implemented by parents, for increasing phonological awareness (PA). Three kindergarten students identified with PA deficits and their mothers served as participants. Parents implemented Sound Foundations twice per week for five and a half weeks. The effects of the intervention were evaluated using a multiple baseline across participants design. PA was measured using the Phoneme Segmentation Fluency measure of the Dynamic Indicators of Basic Early Literacy Skills. Results showed that two students showed increases over time in phonemes per minute after the parent intervention. A third student did not show substantial mprovement. None of the students reached established levels of PA. Treatment integrity and social validity data also were collected. These data indicated that parent participants carried out the intervention appropriately and that both parents and student liked the intervention program.

It has been noticed for a long time that humans can identify others based on their biological movement from a distance. However, it is only recently that computer vision based gait biometrics has received much attention. In this dissertation, we perform a thorough study of gait recognition from a computer vision perspective. We first present a parameterless baseline recognition algorithm, which bases similarity on spatio-temporal correlation that emphasizes gait dynamics as well as gait shapes. Our experiments are performed with three popular gait databases: the USF/NIST HumanID Gait Challenge outdoor database with 122 subjects, the UMD outdoor database with 55 subjects, and the CMU Mobo indoor database with 25 subjects. Despite its simplicity, the baseline algorithm shows strong recognition power. On the other hand, the outcome suggests that changes in surface and time have strong impact on recognition with significant drop in performance. To gain insight into the effects of image segmentation on recognition -- a possible cause for performance degradation, we propose a silhouette reconstruction method based on a Population Hidden Markov Model (pHMM), which models gait over one cycle, coupled with an Eigen-stance model utilizing the Principle Component Analysis (PCA) of the silhouette shapes. Both models are built from a set of manually created silhouettes of 71 subjects. Given a sequence of machine segmented silhouettes, each frame is matched into a stance by pHMM using the Viterbi algorithm, and then is projected into and reconstructed by the Eigen-stance model. We demonstrate that the system dramatically improves the silhouette quality. Nonetheless, it does little help for recognition, indicating that segmentation is not the key factor of the covariate impacts. To improve performance, we look into other aspects. Toward this end, we propose three recognition algorithms: (i) an averaged silhouette based algorithm that deemphasizes gait dynamics, which substantially reduces computation time but achieves similar recognition power with the baseline algorithm; (ii) an algorithm that normalizes gait dynamics using pHMM and then uses Euclidean distance between corresponding selected stances -- this improves recognition over surface and time; and (iii) an algorithm that also performs gait dynamics normalization using pHMM, but instead of Euclidean distances, we consider distances in shape space based on the Linear Discriminant Analysis (LDA) and consider measures that are invariant to morphological deformation of silhouettes. This algorithm statistically improves the recognition over all covariates. Compared with the best reported algorithm to date, it improves the top-rank identification rate (gallery size: 122 subjects) for comparison across hard covariates: briefcase, surface type and time, by 22%, 14%, and 12% respectively. In addition to better gait algorithms, we also study multi-biometrics combination to improve outdoor biometric performance, specifically, fusing with face data. We choose outdoor face recognition, a "known" hard problem in face biometrics, and test four combination schemes: score sum, Bayesian rule, confidence score sum, and rank sum. We find that the recognition power after combination is significantly stronger although individual biometrics are weak, suggesting another effective approach to improve biometric recognition. The fundamental contributions of this work include (i) establishing the "hard" problems for gait recognition involving comparison across time, surface, and briefcase carrying conditions, (ii) revealing that their impacts cannot be explained by silhouette segmentation, (iii) demonstrating that gait shape is more important than gait dynamics in recognition, and (iv) proposing a novel gait algorithm that outperforms other gait algorithms to date.

This study investigated the efficacy of multiple indicators, multiple causes (MIMIC) methods in detecting uniform and nonuniform differential item functioning (DIF) among multiple groups, where the underlying causes of DIF was different. Three different implementations of MIMIC DIF detection were studied: sequential free baseline, free baseline, and constrained baseline. In addition, the robustness of the MIMIC methods against the violation of its assumption, equal factor variance across comparison groups, was investigated. We found that the sequential-free baseline methods provided similar Type I error and power rates to the free baseline method with a designated anchor, and much better Type I error and power rates than the constrained baseline method across four groups, resulting from the co-occurrence background variables. But, when the equal factor variance assumption was violated, the MIMIC methods yielded the inflated Type I error. Also, the MIMIC procedure had problems correctly identifying the sources DIF, so further methodological developments are needed.

This thesis details work undertaken in the field of radio astronomy instrumentation. Specific components of the data collection and processing systems used by radio astronomers have been implemented using non-traditional approaches. Traditionally, the correlation of radio astronomy data has taken place on dedicated, specific hardware. This thesis deals with the implementation of equivalent correlators using software running on generic clusters of personal computers - the software approach to radio astronomy. Toward this end a system has been developed that records the raw telescope output onto computer hard drives, allowing easy access to the data on cluster supercomputers. Part of this thesis describes the design, construction, testing and utilisation of these data recording systems. The correlator software used to process these data on supercomputers is also fully described, including extensive tests of the software and a detailed comparison between its output and the output of an existing hardware correlator. The software correlator is shown to produce output that agrees extremely well with the hardware correlator, verifying its accuracy and performance. Finally, results of on-going scientific investigations that use the software correlators described in this thesis are outlined, illustrating the flexibility and usefulness of the software approach to radio astronomy.

The self-calibration method, a new interferometric technique using measurements in the K′-band (2.1 μm) at the CHARA Array, has been used to derive orbits for several spectroscopic binaries. This method uses the wide component of a hierarchical triple system to calibrate visibility measurements of the triple’s close binary system through quasi-simultaneous observations of the separated fringe packets of both. Prior to the onset of this project, the reduction of separated fringe packet data had never included the goal of deriving visibilities for both fringe packets, so new data reduction software has been written. Visibilities obtained with separated fringe packet data for the target close binary are run through both Monte Carlo simulations and grid search programs in order to determine the best-fit orbital elements of the close binary. Several targets, with spectral types ranging from O to G and luminosity classesfrom III to V, have been observed in this fashion, and orbits have been derived for the close binaries of eight targets (V819 Her B, Kappa Peg B, Eta Vir A, Eta Ori Aab, 55 UMa A, 13 Ceti A, CHARA 96 Ab, HD 129132 Aa). The derivation of an orbit has allowed for the calculation of the masses of the components in these systems. The magnitude differences between the components can also be derived, provided that the components of the close binary have a magnitude difference of Delta K < 2.5 (CHARA’s limit). Derivation of the orbit also allows for the calculation of the mutual inclination (Phi), which is the angle between the planes of the wide and close orbits. According to data from the Multiple Star Catalog, there are 34 triple systems other than the 8 studied here for which the wide and close systems both have visual orbits. Early formation scenarios for multiple systems predict coplanarity (Phi < 15 degrees), but only 6 of these 42 systems are possibly coplanar. This tendency against coplanarity may suggest that the capture method of multiple system formation is more important than previously believed.

This thesis presents an engineering project measuring the effective wavelength of the CHARA Classic beam combiner on the CHARA Array. Knowing the actual effective wavelength of light observed is very important because that value is necessary for determining astrophysical parameters of stars. Currently, the value used for CHARA Classic data comes from a model of the system and is based on numbers published by the manufacturer of the filter; it is not derived from measurements done on the system directly. We use two data collection methods to observe standard stars of different spectral types and calculate the wavelength of light recorded by the instrument for each star. We find the best estimate of the effective wavelength for the CHARA Classic K′-band configuration to be 2.138±0.003μm, a 0.56% decrease from the previously adopted value of 2.150μm. Our result establishes the first estimate of the uncertainty in the effective wavelength.

Spinal fusion is a surgical procedure used to correct structural spinal damage or abnormalities. Recovery is painful and consists of a minimum 3-day hospital stay. Specific body positioning is necessary for healing but is difficult to maintain due to physical discomfort. The purpose of this study was to use a single-subject multiple baseline design to compare the current practice of using standard hospital pillows to a body-sized pillow for increasing comfort and decreasing pain in pediatric patients recovering from spinal fusion surgery. Four adolescents who had recently undergone spinal fusion surgery served as participants. Outcome measures included self- and nurse-report, heart rate, and requested medication. Three patients found that the BodyPillow® increased their comfort as they recovered from surgery; the fourth reported that he was less comfortable. No changes in pain were reported with the BodyPillow®. Results should help guide medical care and future research regarding pediatric spinal fusion recovery.

ECG measures electrical potentials on the body surface via contact electrodes. Conditions such as movement of the patient, breathing, and interaction between the electrodes and skin cause baseline wandering of the ECG signal. Baseline wandering noise can mask some important features of the ECG signal / hence it is desirable to remove this noise for proper analysis of the ECG signal. This study includes an implementation and evaluation of methods to remove this noise, such as finite impulse response filters, infinite impulse response filters, interpolation filters and adaptive filters. These filters are first applied offline to simulated ECG data. The filter outputs and their frequency spectra are compared to the pure ECG signal and its frequency spectrum using visual inspection and numerical evaluation criteria such as root mean squared error and percentage root relative squared error. The best filters are then selected and applied online to the same simulated data. Finally, these best methods are used to suppress the baseline wandering noise in real ECG recordings using both offline and online filtering. In the offline application, windowing type filters were found to be the most successful filters among the implemented filters. However, a high filter order should be used to produce such good results, which increases the computation time, thus it may not be the best method for online filtering, in which fast computation is essential. Butterworth bidirectional type is preferred for online filtering since it has lower computational complexity, and it produces acceptable results.