Evaluation of organizational effectiveness in a management

Rowley, Thomas Howard 1987 국내박사

The new humanism in the thratre of Edward Albee

Shull, Kathleen Rowley U. M. I. 1983 국내박사

Neural network-based face detection

Rowley, Henry Allan Carnegie-Mellon University 1999 해외박사(DDOD)

Object detection is a fundamental problem in computer vision. For such applications as image indexing, simply knowing the presence or absence of an object is useful. Detection of faces, in particular, is a critical part of face recognition and, and critical for systems which interact with users visually. Techniques for addressing the object detection problem include those matching a two- and three- dimensional geometric models to images, and those using a collection of two-dimensional images of the object for matching. This dissertation will show that the latter view-based approach can be effectively implemented using artificial neural networks, allowing the detection of upright, tilted, and non-frontal faces in cluttered images. In developing a view-based object detector using machine learning, three main subproblems arise. First, images of objects such as faces vary considerably with lighting, occlusion, pose, facial expression, and identity. When possible, the detection algorithm should explicitly compensate for these sources of variation, leaving as little as possible unmodelled variation to be learned. Second, one or more neural networks must be trained to deal with all remaining variation in distinguishing objects from non-objects. Third, the outputs from multiple detectors must be combined into a single decision about the presence of an object. This thesis introduces some solutions to these subproblems for the face detection domain. A neural network first estimates the orientation of any potential face. The image is then rotated to an upright orientation and preprocessed to improve contrast, reducing its variability. Next, the image is fed to a frontal, half profile, or full profile face detection network. Supervised training of these networks requires examples of faces and nonfaces. Face examples are generated by automatically aligning labelled face images to one another. Nonfaces are collected by an active learning algorithm, which adds false detections into the training set as training progresses. Arbitration between multiple networks and heuristics, such as the fact that faces rarely overlap in images, improve the accuracy. Use of fast candidate face selection, skin color detection, and change detection allows the upright and tilted detectors to run fast enough for interactive demonstrations, at the cost of slightly lower detection rates. The system has been evaluated on several large sets of grayscale test images, which contain faces of different orientations against cluttered backgrounds. On their respective test sets, the upright frontal detector finds 86.0% of 507 faces, the tilted frontal detector finds 85.7% of 223 faces, and the non-frontal detector finds 56.2% of 96 faces. The differing detection rates reflect the relative difficulty of these problems. Comparisons with several other state-of-the-art upright frontal face detection systems will be presented, showing that our system has comparable accuracy. The system has been used successfully in the Informedia video indexing and retrieval system, the Minerva robotic museum tour-guide, the WebSeer image search engine for the YAW, and the Magic Morphin' Mirror interactive video system.

Mechanisms of interfacial electron-transfer within high-surface-area metal-oxide thin films

Rowley, John G The Johns Hopkins University 2011 해외박사(DDOD)

The direct conversion of solar photon energy into electrical power is achieved with photovoltaic technology, yet existing technology is too inefficient or expensive to implement on a global scale. Dye sensitized solar cells (DSSCs) based on earth abundant low cost materials could overcome the barriers for world-wide implantation of photovoltaic technology. Historically the most efficient regenerative DSSCs utilize iodide based redox mediators in nitrile solvents. Despite the dominance of iodide redox mediator in DSSCs, the chemical attribute(s) that make iodide based electrolytes superior to other electrolyte is yet unproven. Discovering the chemical cause of iodide's superiority as a redox mediator was and important aspect of this thesis research. In Chapter 2 the photoinitated formation and cleavage of I-I bonds is demonstrated at an un-sensitized potentiostatically controlled mesoporous nanocrystalline (anatase) TiO2 thin film. After pulsed laser excitation of a TiO2 thin film, I2·- was observed that disproportionated to yield I3- and I-, but did not react with TiO2. In contrast evidence for a quasi-Fermi level dependent reaction between TiO2(e -) and I3- was observed. In agreement with the findings at un-sensitized TiO2 in Chapter 2, the results presented in Chapter 3 demonstrate that after pulsed laser excitation of an operational DSSC: I2·- is observed, I2·-does not react with TiO 2, and that charge recombination between I3- and TiO2(e-) is operative. In addition to absorption changes attributed to iodide redox chemistry, evidence for a transient electric field induced Stark effect was observed in the operational DSSC. Absorption changes reporting on the electric field at the TiO2 surface were simultaneously quantified at specific power conditions fo the operational DSSC. Chapter 4 explores alternative redox mediators to the iodide/triiodide couple. The photophysical studies presented herein provide key mechanistic details on kinetic processes pertinent to operational DSSCs. Specifically, electron transfer to and from iodide species found in an operational DSSC were quantified. Taken together this research advances the body of knowledge quantifying the chemical properties that make iodide based electrolytes a superior redox mediator in dye sensitized solar cells.

Women's rights and women's work: Politics and professionalism in 19th century American fiction (Sarah J. Hale, Laura Curtis Bullard, Harriet Beecher Stowe, Louisa May Alcott, Elizabeth Stuart Phelps)

Rowley, Cathleen J City University of New York 2005 해외박사(DDOD)

Throughout the nineteenth-century, debates on women's work abounded in American fiction and in the press. While much literature and most advice manuals for women were inclined towards extolling the virtues of domesticity, there exists a body of work devoted to a different viewpoint---namely, that women were working outside of the home and that these women needed guidance, encouragement, and practical advice. While the literature written by American women in the mid-nineteenth century is often characterized as domestic or sentimental fiction, literature during this time also provided critiques of contemporary society that were decidedly unsentimental and concerned with women's movement into the public sphere. This dissertation examines representations of working women in American fiction from 1820--1885, specifically focusing on the rise or the development of the professional woman. In charting the increasing presence of the professional woman in American fiction, I explore the relationships between the women's rights movement and fiction which contained feminist characters. Examining literary works from 1820--1885, I trace the progression of the professional woman from early representations of women's rights lecturers and schoolteachers to later portrayals of professionals such as writers, artists, and physicians. The authors considered are both male and female, and range from canonical writers to little-known or neglected authors, including Sarah J. Hale, Laura Curtis Bullard, Harriet Beecher Stowe, Louisa May Alcott, and Elizabeth Stuart Phelps among others. I argue that the figure of the professional working woman in literary works of the time reflects working women of the period, but at the same time, also critiques elements of American culture, particularly in relation to restrictions on women's lives, both in domestic matters and in professional capacities outside of the home.

The elastic properties of the eyewall, the effect of focal pressure on the retina, and the development of reversible bioadhesives for utilization in a retinal prosthesis

Rowley, Adrian Paul University of Southern California 2011 해외박사(DDOD)

Degenerative retinal diseases such as Retinitis Pigmentosa and Age-Related Macular Degeneration are leading causes of blindness. Retinal prostheses have been developed to bypass the degenerate photoreceptors and stimulate the viable bipolar and retinal ganglion cells. These prostheses are designed to apply electrical stimulation directly to the viable cells, creating an artificial version of sight in implanted patients. Epiretinal prostheses utilize a polymer electrode array, which is affixed to the inner limiting membrane of the retina, using a retinal tack. The electrode array and tack creates varying focal pressure across the interface and subsequent retinal damage. In order to assist bioengineers to make design changes that can minimize these effects, firstly the elastic properties of the eyewall tissues must be accurately established. These properties can be utilized in computerized modeling (Finite Element Analysis) of the interface and enable refinement of future array designs. Secondly, defining safe limits of the amount of focal pressure that the retina can withstand without significant pathological damage is equally important. Thirdly, the integration of reversible bioadhesives could allow intraocular arrays to maintain intimate contact with the retina and may eliminate the need for a retinal tack. I have defined the elastic properties of the porcine and human posterior ocular poles including the retina, choroid and sclera in both the vertical and horizontal meridian. I describe the biological effect of focal pressure on the retina and attempt to define safe limits for future designs. I have also developed and tested reversible bioadhesives that could be integrated into the intraocular electrode array, in order to eliminate the tack. We have implanted these adhesive biomaterials onto the retina of both rabbits and dogs. These bioadhesives may have multiple applications not only in ophthalmology, but in a wide variety of surgical implants. The potential benefits for the retinal prosthesis could include removing the need for the tissue destructive tack, a reduction in focal pressure effects, improved apposition of the intraocular array to the retina, minimization of the current required to stimulate neurons, extended battery life, diminished heat production, and ultimately improved visual acuity.

Antiviral natural products from marine sources

Rowley, David Chapman University of California, San Diego 2001 해외박사(DDOD)

Viruses have probably afflicted mankind for millennia, but modern factors such as the booming human population, fast and widespread global travel, and increased ecosystem perturbations have assured viral disease of an even brighter spotlight in the near future. The examples are frightening. HIV infection is still spreading around the globe. Deadly outbreaks of viruses such as Ebola and Lassa fever continue to occur ever more closely to heavily populated regions. Drug resistant strains of familiar pathogens such as herpes simplex virus are beginning to emerge. Clearly, our science is challenged to find new treatments for these diseases. How will the next generation of antiviral medicines be discovered? A crucial component should continue to be the proven strategy of screening natural products. In the past, structurally diverse metabolites produced by terrestrial microorganisms and plants have shown inhibitory activities against various viral diseases. Unlike their terrestrial counterparts, however, the antiviral potential of secondary metabolites produced by marine organisms is virtually unexplored. This dissertation describes the discovery of novel antiviral agents produced by marine organisms. Studies involving several important viral targets are presented, and included are chemical investigations that have identified lead compounds for antiviral drug discovery. A series of linear, lipophilic peptides produced by a marine fungus are presented as novel inhibitors of the herpes simplex virus. Evidence suggests that these peptides directly inactivate the virus. Synthetic analogs of the natural substrates were prepared, leading to the development of structure-activity relationships outlining the biological importance of key structural features, and perhaps a greater insight into the possible mechanism of action. The thalassiolins, metabolites isolated from a sea grass <italic>Thalassia testudinum</italic>, were identified as inhibitors of the HIV enzyme integrase, and also the replication of HIV in cell culture. The marine cyclic peptide sansalvamide is presented as the first natural product inhibitor of the topoisomerase enzyme of the Molluscum contagiosum virus. I believe that the molecules described in this dissertation extend our knowledge and comprehension of viral inactivating agents, and further serve to support the recognition of the marine environment as a future resource for the continued discovery of novel antiviral medicines.

Controlling myoblast phenotype with RGD-modified alginate matrices

Rowley, Jon A University of Michigan 2001 해외박사(DDOD)

New strategies are being developed to grow tissues and organs for transplantation from cells and biomaterials. Biomaterials are critical components of engineered tissues, as they act as scaffolds for new tissues development. To date, however, engineered tissues do not have the complete structure or function of native tissues they are to replace. It is important to provide cells within these tissues with biological signals to help guide cell function during engineered tissue development. We hypothesized that the phenotype of cells adherent to a biomaterial could be regulated by controlling the mechanism of cell adhesion to the biomaterial, and we developed a model biomaterial system based on alginate hydrogels to address this hypothesis. Alginates are hydrophilic polysaccharides composed of mannuronic (M) and guluronic (G) acid monomers that gel in the presence of divalent cations such as Ca<super>+2</super>. We modified the alginates with RGD-peptides using carbodiimide chemistry. Ligand type and density may be varied on alginates of varying M:G over several orders of magnitude, with incorporation efficiency typically >60%. Myoblasts adhered, proliferated and differentiated on RGD-alginate hydrogels with a surface density of 10 fmols/cm2. Myoblast adhesion specificity was demonstrated, as soluble RGD (1mM) completely inhibited adhesion to the substrates, while RGE-peptides (1mM) had no effect. Myoblast function was controlled by varying ligand type, ligand density, and M:G of the alginate. Myoblast proliferation increased on RGD vs. YIGSR peptide ligands (density = 10 fmols/cm<super>2</super>), and proliferation and fusion increased by increasing RGD-density from 1–100 fmols/cm<super>2</super>. Furthermore, myoblast proliferation and fusion, but not muscle-specific gene expression, were dependent on the M:G ratio of alginates. Varying M:G ratio from 30:70 to 65:35 increased myoblast proliferation and fusion, but did not alter the expression of MyoD or myogenin (myoblast transcription factors) or the activity levels of creatine kinase. This effect was caused by calcium calcium ions released from different alginate substrates. In conclusion, myoblast phenotype may be controlled by varying ligand type and density at the material surface, and by modulating local calcium concentrations using different alginate types.

Comparison of Variable Selection Methods

Rowley, Elizabeth Koehler ProQuest Dissertations & Theses The University of 2019 해외박사(DDOD)

Use of classic variable selection methods in public health research is quite common. Many criteria, and various strategies for applying them, now exist including forward selection, backward elimination, stepwise selection, best-subset selection and so on, but all suffer from similar drawbacks. Chief among them is a failure to account for the uncertainty contained in the model selection process. Ignoring model uncertainty can cause several serious problems. Variance estimates are generally underestimated, p-values are generally inflated, prediction ability is overestimated, and results are not reproducible in another dataset. Modern variable selection methods have become increasingly popular, especially in applications of high-dimensional or sparse data. Some of these methods were developed to address the short-comings of classic variable selection methods, such as backward elimination and stepwise selection methods. However, it remains unclear how modern variable selection methods behave in a classical, meaning non-high-dimensional, setting. A simulation study investigates the estimation, predictive performance and variable selection capabilities of three representative modern variable selection methods: Bayesian model averaging (BMA), stochastic search variable selection (SSVS), and the adaptive lasso. These three methods are considered in the setting of linear regression with a single variable of interest which is always included in the model. A second simulation study compares BMA to classical variable selection methods, including backward elimination, two-stage method, and change-in-effect method in the setting of logistic regression. Additionally, the data generated in both simulation studies closely mimic a real study and reflect a realistic correlation structure between potential covariates. Sample sizes ranging from 150 to 20000 are investigated. BMA is demonstrated in an example building a predictive model using data from the China Health and Nutrition Survey.

WHEN SPIRIT ENTERS THE CORPORATE BODY: SHEPHERDING THE CORPORATE SOUL THROUGH TRANSFORMATION

ROWLEY, THERESE M THE UNION INSTITUTE 1997 국내박사