No description available.

No description available.

Evaluation of diel and seasonal patterns in offshore marine mammal activity through visual data collection can be impaired by poor weather and light limitations and by the requirement for costly ship time. As a result, relatively little is known about the diel patterns of wild dolphins. Pacific white-sided dolphins north of Southern California are particularly under-researched. Collecting acoustic data can be a cost-effective approach to evaluating activity patterns in offshore marine mammals. However, manual analysis of acoustic data is time-consuming, and impractical for large data sets. This study evaluates diel and seasonal patterns in Pacific white-sided dolphin communication through automated analysis of one year of continuous acoustic data collected from the Barkley Canyon node of Ocean Networks Canada’s NEPTUNE observatory, offshore Vancouver Island, British Columbia, Canada. In this study, marine mammal acoustic signals are manually annotated in a sub-set of the data, and used to train a random forest classifier targeting Pacific white-sided dolphin pulsed calls. Marine mammal vocalizations are classified using the resultant classifier, manually verified, and examined for seasonal and diel patterns. Pacific white-sided dolphins are shown to be vocally active during all diel periods in the spring and summer, but primarily at dusk and night in the fall and winter. Additionally, the percentage of time they are detected drops significantly in the fall and remains low during the winter. This pattern suggests that a group of day-active dolphins, possibly a unique population, leaves Barkley Canyon in the fall and returns in the spring. It is hypothesized that this group may be following the Pacific herring, which are present at the surface during the day at Barkley Canyon in the spring and summer, and migrate inshore for the fall and winter. / Graduate

No description available.

No description available.

No description available.

Acoustic Emission Analysis (AEA) is the study of the sonic (and ultrasonic) energy released by chemical systems in the form of transient waves, as the system attempts to (re)attain equilibrium. This area of chemistry, and chemical analysis, is ripe for fundamental studies since it has been little explored. The high potential of the technique as a non-invasive, non-destructive reaction monitoring scheme suggests that numerous applications will follow. In this work, an apparatus and software have been constructed to monitor acoustic emission (AE) and collect and process AE data. A broad-band piezoelectric transducer was used to convert the acoustic signals to electrical waveforms which could be captured by a digital storage oscilloscope. These waveforms were then stored on an IBM-compatible computer for further analysis. Analysis of the data was performed using pattern recognition techniques. The signals were characterized through the use of descriptors which can map each signal onto a multi-dimensional feature space. Visualization of the data structure in multidimensional space was accomplished using several methods. Hierarchical clustering was used to produce tree structures, known as dendrograms, which attempt to show clustering of the signals into various groups. Abstract factor analysis (AFA) - also called principal components analysis (PCA) - was used to project the data onto a two dimensional factor space to allow for direct viewing of structure in the multidimensional data. Sodium hydroxide dissolution, aluminum chloride hydration and heat activation of Intumescent Flame Retardants (IFR's) were used to test the assembled hardware and to provide data to submit to the pattern recognition algorithms coded as part of this work. The solid-solid phase transition of trimethylolethane (Trimet), and the liquid crystal phase transitions of two liquid crystals (α-ѡ-bis(4-n-decylaniline-benzilidene-4'-oxyhexane), and 4-n-pentyloxybenzylidene-4'-n-heptylaniline) were also monitored and the signals analyzed. The pattern recognition software was able to extract much information from the acoustically emitting samples - information which would not have been apparent by using standard (uni- and bi-variate) methods of analysis. Chemical acoustic emission, coupled with pattern recognition analysis, will be able to provide the chemist with knowledge (qualitative, quantitative, kinetic, etc.) about chemical systems which are often difficult or impossible to monitor and analyze by other means. / Science, Faculty of / Chemistry, Department of / Graduate

archives@tulane.edu / Acoustic droplet vaporization (ADV) is an attractive alternative to traditional hepatocellular carcinoma (HCC) treatments. ADV involves injecting microdroplets into the bloodstream which then accumulate in and around the tumor’s vasculature. Once accumulated, high-power ultrasound is used to vaporize the microdroplets into larger perfluoropentane gas microbubbles which occlude blood flow and induce necrosis of the tumor without harming healthy tissue like traditional HCC treatments. This study aims to optimize ADV treatment by improving the shell composition and surface architecture of microdroplets while ensuring the treatment remains safe. In order to ensure the treatment is as effective as possible, the microdroplets must have powerful binding capabilities, guaranteeing maximum microdroplet accumulation and treatment efficacy. The binding capabilities of three microdroplet shell compositions, created by adjusting the molar percentages of the three lipids found in the shell, were investigated and found to all have equal binding abilities. The surface architecture of these microdroplets were also altered to maximise binding capabilities. Microdroplets can have either an exposed-ligand or buried-ligand surface architecture. In microdroplets with a buried-ligand surface architecture, the attached tumor-targeting ligands are hidden within a layer of longer lipid chains which allow the microdroplets to evade the immune system and circulate within the bloodstream longer, increasing treatment efficacy. It was found that microdroplets with a buried-ligand surface architecture do not have comparable binding capabilities to microdroplets with an exposed-ligand surface architecture and are therefore not a viable alternative for use in ADV. Finally, the velocity required to dislodge perfluoropentane gas microbubbles was explored to determine if the gas microbubbles can remain adhered to the tumor’s vasculature to create a strong occlusion. Since perfluoropentane gas microbubbles occlude blood flow it is imperative that the microbubbles remain in the tumor’s vasculature and do not dislodge and accumulate in other parts of the body’s vasculature. By measuring the velocity and calculating the force necessary for dislodgement and comparing those values to those found in capillaries it was concluded that the perfluoropentane gas microbubbles can withstand the force of blood flow and remain lodged in capillaries. / 1 / Chloe Celingant-Copie

In the past decades, statistics-based hidden Markov models (HMMs) have become the predominant approach to speech recognition. Under this framework, the speech signal is modeled as a piecewise stationary signal (typically over an interval of 10 milliseconds). Speech features are assumed to be temporally uncorrelated. While these simplifications have enabled tremendous advances in speech processing systems, for the past several years progress on the core statistical models has stagnated. Since machine performance still significantly lags human performance, especially in noisy environments, researchers have been looking beyond the traditional HMM approach. Recent theoretical and experimental studies suggest that exploiting frame-torame correlations in a speech signal further improves the performance of ASR systems. This is typically accomplished by developing an acoustic model which includes higher order statistics or trajectories. Linear Dynamic Models (LDMs) have generated significant interest in recent years due to their ability to model higher order statistics. LDMs use a state space-like formulation that explicitly models the evolution of hidden states using an autoregressive process. This smoothed trajectory model allows the system to better track the speech dynamics in noisy environments. In this dissertation, we develop a hybrid HMM/LDM speech recognizer that effectively integrates these two powerful technologies. This hybrid system is capable of handling large recognition tasks, is robust to noise-corrupted speech data and mitigates the ill-effects of mismatched training and evaluation conditions. This two-pass system leverages the temporal modeling and N-best list generation capabilities of the traditional HMM architecture in a first pass analysis. In the second pass, candidate sentence hypotheses are re-ranked using a phone-based LDM model. The Wall Street Journal (WSJ0) derived Aurora-4 large vocabulary corpus was chosen as the training and evaluation dataset. This corpus is a well-established LVCSR benchmark with six different noisy conditions. The implementation and evaluation of the proposed hybrid HMM/LDM speech recognizer is the major contribution of this dissertation.

No description available.