Volume holographic memory (VHM) systems are page-oriented optical data storage systems with a relatively large storage capacity, a high aggregate data transfer rate, and a fast access time as compared with other conventional mass data storage systems. In addition to the noise sources commonly present in a conventional communication channel, the VHM systems use a two-dimensional (2-D) data format and therefore suffers from 2-D electrical/optical hybrid noises. In this dissertation, we apply digital communication coding techniques to such a page-oriented optical memory system. To establish a coding performance baseline, we take an information theoretical perspective to study the design of 4F optical imaging systems, and the storage capacity of page-wise accessed VHM systems. In addition to the information-based figure of merit, bit-error rate (BER) serves as a more practical data fidelity measure. A BER-based optical design and a constant BER recording schedule in VHM systems are shown to support the coding analyses based on the BER metric. We analyze the VHM storage capacity while applying Reed-Solomon error-correction coding (ECC) and demonstrate that the use of the 2-D interleaving together with ECC results in a significant capacity improvement in the presence of systematic errors. In addition, soft-decision iterative decoding affords us a 2-D parallel decoding scheme that unifies the 2-D page equalization and decoding. A highly parallel decoder is discussed in the context of the 2-D low-pass channel mitigation and error correction. In the last part of this dissertation, we consider hardware implementation issues in 2-D optical systems. We design two coding related components using the MOSIS Orbit 2 mum n-well process to fabricate two CMOS VLSI chips. Power consumption and aerial scaling laws are discussed for each applications.

Ces dernières années, l'élastographie ultrasonore et IRM s'est développée afin de caractériser les propriétés viscoélastiques des tissus mous. Ces nouvelles méthodes d'imagerie sont particulièrement prometteuses pour caractériser des pathologies comme les carcinomes dans le sein qui présentent une élasticité plus importante que les tissus environnants. Dans la première partie du manuscrit, une présentation des milieux viscoélastiques permet de décrire les comportements mécaniques observés lorsqu'un tel milieu est sollicité par une onde élastique. Différents modèles de propagation d'ondes de cisaillement sont analysés afin d'arrêter un choix quant au modèle utilisé pour estimer les propriétés mécaniques des tissus biologiques par élastographie IRM. Dans la seconde partie, les différentes méthodes d'élastographie par imagerie ultrasonore sont présentées avant d'introduire quelques bases de l'imagerie par résonance magnétique nécessaires à la compréhension de l'élastographie IRM. Deux approches de programmation du codage des déplacements sont comparées théoriquement et expérimentalement de manière à optimiser la qualité de l'estimation des propriétés mécaniques. Enfin la troisième partie aborde les difficultés de mise en oeuvre de ces deux techniques d'acquisition pour certains organes du corps humain comme la carotide ou le coeur, et propose une nouvelle approche de l'élastographie IRM pour pallier ces problèmes. Les bases théoriques de cette méthode sont présentées et comparées aux techniques conventionnelles d'élastographie IRM. Des expériences dans un gel permettent alors de confirmer les observations théoriques réalisées. Enfin, cette technique d'acquisition est adaptée aux contraintes de l'élastographie du coeur et le suivi des propriétés mécaniques in vivo du coeur durant le cycle cardiaque sont présentés sous forme d'élastogrammes.

non fournis

Non Disponible

Techniques for implementing the EM and simulated-annealing image reconstruction algorithms on a large-grain parallel computer for faster execution per iteration are developed for emission tomography applications. The speedups obtained by implementing the algorithms on up to 54 processors connected in a ring topology are found to be nearly linear. Reconstruction involves finding an estimate of the emission distribution that minimizes an energy function that contains a data-agreement term and a noise-control term. The EM algorithm minimizes the complete-incomplete form of the data-agreement term, which is easily partitioned for parallel computation. The simulated-annealing algorithm is a Monte Carlo method in which any form of data-agreement and noise-control term can be minimized. In the reconstruction of a thyroid phantom, it is demonstrated that the complete-incomplete data-agreement term can be used to facilitate the parallel implementation of simulated annealing while still guaranteeing convergence.

Pattern recognition problems involve two main issues: feature formulation and classifier design. This thesis is concerned with the latter. Numerous algorithms for the design of pattern recognition systems have been published, and the algorithm detailed herein is a new approach--specific to the design of decision tree classifiers. It involves a top-down strategy, optimizing the root node decision and then subsequently its children. To assess various pattern space partitions, the Tie statistical distance measure quantified the separability of potential cluster groupings. Additionally, a separate neural network was employed at each of the tree decision nodes. Results from the application of this methodology to the regional labeling of panchromatic images suggest it is a suitable approach.

Conventional non-contact optical methods of surface profiling are limited in the range of surface heights that can be accurately measured due to phase ambiguity errors on steep local slopes. Instruments that have been developed thus far to avoid the problems with local slope typically suffer from poor measurement height resolution and slow measurement speeds. Contact profilometers such as stylus-based instruments suffer from poor lateral resolution due to the finite radius of the stylus tip, and slow measurement speeds, especially when two-dimensional scans of the surface are required. Stylus tips can also scratch delicate surfaces during the measurement. We propose a new method of optical, non-contact profiling of rough surfaces without the limitations on local slope that other methods suffer from. This new method utilizes interferometric techniques as well as digital signal processing algorithms to produce fast, accurate, and repeatable three-dimensional surface profile measurements on a wide variety of surfaces.

Since the discovery of a class of superconducting materials with critical temperatures as high as 125 degrees kelvin, there has been a great deal of research interest in their possible application to optical radiation detection, particularly in the infrared spectrum. The motivation for this research is the promise of a fast detector operating at elevated temperatures that is sensitive to low level optical signals and that operates out to the far IR. It has been shown that thin films of these high temperature superconductors (HTS) exhibit a change in their electrical properties when exposed to optical radiation. However, in order to make a practical detector out of HTS materials, the mechanisms of this response must be fully understood. The purpose of this research is to investigate the spectral, temporal and thermal characteristics of this electrical response in an effort to better understand the mechanisms involved.

It is a common misconception to consider the intensified charge coupled device (ICCD) to be simply a finite element detector. In reality, one must consider the ICCD as an optical system in and of itself, complete with its own throughput and point spread function. Only in this light can an accurate model for the ICCD performance be developed. Variations in signal to noise ratio (SNR) with image size can have broad implications in the design and modeling of ICCD based systems. Results from an experiment devised to measure the variation of SNR performance with input image size for selected Generation 2 microchannel plate (MCP) ICCD's are presented. Also presented is a generalized theoretical model for ICCD SNR performance for small images.

When used in conjunction with an optical system, a nutating reticle can be utilized to track point sources. The position information of a single point source in the field of view is encoded as a phase modulated signal that is demodulated by appropriate electronics to extract the angular position of the single point source for tracking purposes. With two point sources in the field of view, it can be shown that particular types of demodulation electronics will output position information such that the tracking system will track the irradiance weighted centroid of the two sources.