Doctor of Philosophy / The digital imaging technology, geographical analyses tool, and computer vision (a technique that enables computers and systems to get meaningful information from images) methods can be used to extract traits-related branching pattern, canopy cover, and pod location in edamame for many plant populations in short time using less labor and resources. Using genome-wide association study, we identified several genetic markers that were associated with those traits. These markers can be used in marker-assisted selection to develop the edamame varieties that are more adaptable to mechanical harvesting and give more yield, along with understanding the physiological mechanisms for better shoot architecture traits and better yield. We used spectral signatures of different edamame at several harvesting time along with machine learning methods to identify the optimal harvest time of edamame. Hyperspectral imaging (a technique that analyzes a wide spectrum of light instead of just assigning primary colors (red, green, blue) to each pixel) when combined with computer vision and machine learning methods can be used to quantify the levels of vomitoxin (chemical that causes vomiting and feed refusal in animal and humans) for larger wheat kernel samples in a cheaper and faster way.

Manual fracture mapping in tunnels, caverns, mines or other underground spaces is a time intensive and sometimes dangerous process. A system that can automate this task could minimize human exposure to rockfalls, rockbursts or instabilities and facilitate the use of new methods of data visualization such as virtual environments. This research was undertaken to develop VTtrace; a semi-automatic fracture mapping algorithm based on image processing and analysis techniques. Images of a rock exposure surface are made using a "prosumer" grade digital camera. The grayscale images are preprocessed to remove color information and any noise or distortion. The smoothed images are converted into binary images. The binary images are then thinned to extract the fracture map. The fractures are then separated and stored as different images. Fracture properties such as the length, width, orientation and large-scale roughness are determined using photogrammetric techniques. Results from test images shows the VTtrace is effective in extracting rock discontinuity traces. Additional enhancements to the program are proposed to allow feature attributes from the three-dimensional surface to be determined. / Master of Science

Ph.D. university of the Witwatersrand, Faculty of Humanities (Art History), 2012 / Pictorial copies play an essential role in the creation of rock art knowledge, forming a bridge between the art and theories of interpretation. My thesis traces a ‘pictoriography’, that is, a historiography of the practice of recording rock paintings in pictures. I begin with the earliest examples dotting the shifting edges of the Cape Colony from the mideighteenth to mid-nineteenth centuries. Thereafter, the focus shifts to the Maloti-Drakensberg, where two case studies bring this disciplinary history into more recent times. The first is the rainmaking group from Sehonghong Shelter (Lesotho). One of the first rock paintings to be published, it became one of the most iconic in southern Africa. I relate its various copies to one another and to wider views of Sehonghong, revealing how it has been decontextualized and reproduced in diagrammatic form. I develop a ‘digital restoration’, whereby copies circulating independently in the world are returned in digital images to their place of origin. I develop this process further in a site-wide study of eBusingatha Shelter (AmaZizi Traditional Authority Area, KwaZulu-Natal Drakensberg). Once an impressive painted gallery, eBusingatha has been severely damaged by vandalism, removals and collapse, while documents tracking its demise accumulated elsewhere. I reunite scattered records, enabling copies to be contextualized and lost visual qualities of the originals to be restored. Throughout these pictorial genealogies, I explore the distance between the way the rock paintings are illustrated and the way they actually look. While recording strategies are diverse, one dominant convention has emerged in recent decades. Meticulous tracings converted into monochrome redrawings effect a translation of complex and ambiguous painted occurrences into clean forms ‘peeled’ from the rock and projected like shadows onto paper. The are more like text than picture. Colour for instance is considered an integral part of painting traditions worldwide, yet is expunged from the study of San rock paintings. A reintegration of such pictorial attributes into their study may encourage a return to the material world of the imagery and a contextualization of the semantics of its symbolic constituents.

Submitted by Erika Demachki (erikademachki@gmail.com) on 2014-09-22T21:46:54Z No. of bitstreams: 2 Keila Sousa Silva.pdf: 7076994 bytes, checksum: b2ae0f0393e80451a0c468e865f49847 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-09-23T15:39:04Z (GMT) No. of bitstreams: 2 Keila Sousa Silva.pdf: 7076994 bytes, checksum: b2ae0f0393e80451a0c468e865f49847 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-09-23T15:39:04Z (GMT). No. of bitstreams: 2 Keila Sousa Silva.pdf: 7076994 bytes, checksum: b2ae0f0393e80451a0c468e865f49847 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2013-09-23 / This study aims to optimize runtime performance of a system developed to aid the diagnosis of childhood pneumonia by computer. This system, called Pneumocad according Macedo (2012), aims to identify chest radiographs consistent with the disease using computational techniques for recognizing patterns in textures through decomposition of the transformed wavelets, features extracted from the decomposition and classi cation applied to radiographs. In pursuit of this optimization in terms of performance at runtime, insertion of new rays and the recovery of their radiographs similar, we used the proposed deployment of a cluster architecture of radiographs already stored in the database Pneumocad. In parallel , functionality, responsible for de ning how a radiograph is similar to another , were transferred from the source code Java to views in the database. The experiments were performed on databases that contained 183, 2.568 and 10.200 radiographs and using the Pneumo- cad of Macedo (2012) , the Pneumocad Optimized with views and without grouping and Pneumocad Optimized with views and grouping. The experiments and results show that the proposed optimization contributed to the evolution of Pneumocad and enhanced this tool to support the diagnosis of pneumonia in childhood. / O presente estudo propõe a otimizaçao de desempenho de tempo de execu ção de um sistema desenvolvido para auxiliar o diagn ostico de pneumonia infantil por computador. Esse sistema, denominado Pneumocad, segundo Macedo (2012), visa identi car radiografi as de t órax compatí veis com a doen ça utilizando t écnicas de reconhecimento computacional de padrões em texturas por meio da decomposi ção das transformadas wavelets, das caracter sticas extra das das decomposi ções e da classi ca ção aplicadas as radiogra as. Em busca dessa otimiza ção, em termos de desempenho em tempo de execu ção, na inser ção de novas radiogra fias e na recupera ção de suas radiogra fias similares, utilizou-se a proposta da implanta c~ao de uma arquitetura de agrupamento das radiogra fias j a armazenadas na base de dados do Pneumocad. Em paralelo, as funcionalidades, respons aveis por de finir o quanto uma radiogra fia e similar a outra, foram transferidas do c odigo-fonte Java para views na base de dados. Os experimentos foram executados sobre bases de dados que continham 183, 2.568 e 10.200 radiogra fias e utilizando o Pneumocad de Macedo (2012), o Pneumocad Otimizado com views e sem agrupamento e o Pneumocad Otimizado com views e com agrupamento. Os experimentos e resultados mostram que a otimiza c~ao proposta contribuiu para a evolu ção do Pneumocad e aprimorou essa ferramenta que visa apoiar os diagn osticos de pneumonias na infância.

Made available in DSpace on 2016-03-15T19:43:27Z (GMT). No. of bitstreams: 1 Leia Claudia Silva Andrade.pdf: 1697697 bytes, checksum: aef58b06b33c662874778e786364cb73 (MD5) Previous issue date: 2009-02-16 / Fundo Mackenzie de Pesquisa / Contemporary society constantly lives under a flow of images which invades the visual field all the time. Considering this fact, the act of seeing can produce uncertain conditions of interpretations and articulations of the images. Analysing the paintings of Salvador Dalí, Gala Looking to the Mediterranean Sea that, in Eighteen meters away, becomes the portrait of Abraham Lincoln, 1975, will guide us to a path conducted by the act of seeing and also comprehending what has just been seen. This reasearch proposes a deep analysis of the artist Salvador Dalí, as well as dialogue between the fields of art, technology, cultural history and related whenever possible with the pixels of digital images. / A sociedade contemporânea vivencia em situações do cotidiano um universo de imagens que invadem nosso campo visual a todo momento. Nesta trajetória o olhar fragmentado muitas vezes nos dá condições incertas de interpretações e articulações dessas imagens. Neste cenário, a pintura de Salvador Dalí, Gala a Olhar o Mar Mediterrâneo que, a Dezoito Metros de Distância, se Transforma no Retrato de Abraão Lincoln, 1975, nos conduzirá a um itinerário direcionado à prática do ato de olhar e não apenas de ver ou não ver. Este trabalho teórico propõe, dentro deste contexto e da trajetória do artista, um diálogo entre os campos da arte, da tecnologia e história da cultura relacionado sempre que possível com o pixel das imagens digitais.

Com o desenvolvimento das formas de comunicação e expressão ao longo dos séculos, as sociedades puderam experimentar vários modos de fazer arte. Várias finalidades, várias técnicas, algumas tecnologias. A representação da fala pela escrita trouxe conseqüências profundas; entre elas a dominância visual em detrimento dos outros sentidos. Esta nova mídia criou a necessidade de desenvolver linguagens adequadas às suas características. Os computadores possibilitaram a reinterpretação das linguagens existentes e o desenvolvimento de linguagens capazes de integrar todas as outras, com um imenso potencial expressivo. As atuais tecnologias da informação e suas mídias, aliadas à linguagem digital, permitiram uma extrema e radical compressão de tempos e distâncias. A linha de pesquisa deste trabalho está relacionada à experimentação de possibilidades artísticas e expressivas, através do desenvolvimento de estudos visuais com bases conceituais da teoria da Gestalt e da Semiótica. São propostas diferentes utilizações do tempo na releitura digital de obras construtivistas, concretistas e da Optical Art, simulando movimento onde este é sugerido como tensão. A passagem do ponto ao pixel não pode ser feita sem tradução e não deveria ser feita sem interpretação, sem criação. / Through the centuries, human societies have been able to create art in many different ways and developed various communication and expression strategies. Many ends, many techniques, some technologies. The representation of speech through writing brought great results, especially the supremacy of sight at the expense of the other remaining senses. This new media created the need for a customized language to meet its characteristics. Computers made possible a reinterpretation of the existing languages and the development of languages able to integrate all others, with a huge potential for expression. The current information technologies and their media, based on digital languages, led to a radical compression of time and space. This line of research presents some artistic and expressive possibilities through visual studies based on Gestalt and Semiotics general principles. Distinct time interpretations are proposed within the digital version of constructivist, concrete and Op Art artworks, simulating movement where previously it had just been suggested as tension. The shift from dot to pixel is not a translation-free process and should not happen without interpretation, without creation.

Com o desenvolvimento das formas de comunicação e expressão ao longo dos séculos, as sociedades puderam experimentar vários modos de fazer arte. Várias finalidades, várias técnicas, algumas tecnologias. A representação da fala pela escrita trouxe conseqüências profundas; entre elas a dominância visual em detrimento dos outros sentidos. Esta nova mídia criou a necessidade de desenvolver linguagens adequadas às suas características. Os computadores possibilitaram a reinterpretação das linguagens existentes e o desenvolvimento de linguagens capazes de integrar todas as outras, com um imenso potencial expressivo. As atuais tecnologias da informação e suas mídias, aliadas à linguagem digital, permitiram uma extrema e radical compressão de tempos e distâncias. A linha de pesquisa deste trabalho está relacionada à experimentação de possibilidades artísticas e expressivas, através do desenvolvimento de estudos visuais com bases conceituais da teoria da Gestalt e da Semiótica. São propostas diferentes utilizações do tempo na releitura digital de obras construtivistas, concretistas e da Optical Art, simulando movimento onde este é sugerido como tensão. A passagem do ponto ao pixel não pode ser feita sem tradução e não deveria ser feita sem interpretação, sem criação. / Through the centuries, human societies have been able to create art in many different ways and developed various communication and expression strategies. Many ends, many techniques, some technologies. The representation of speech through writing brought great results, especially the supremacy of sight at the expense of the other remaining senses. This new media created the need for a customized language to meet its characteristics. Computers made possible a reinterpretation of the existing languages and the development of languages able to integrate all others, with a huge potential for expression. The current information technologies and their media, based on digital languages, led to a radical compression of time and space. This line of research presents some artistic and expressive possibilities through visual studies based on Gestalt and Semiotics general principles. Distinct time interpretations are proposed within the digital version of constructivist, concrete and Op Art artworks, simulating movement where previously it had just been suggested as tension. The shift from dot to pixel is not a translation-free process and should not happen without interpretation, without creation.

This study links traffic sign visibility and legibility to quantify the effects of damage or deterioration on sign retroreflective performance. In addition, this study proposes GIS-based data integration strategies to obtain and extract climate, location, and emission data for in-service traffic signs. The proposed data integration strategy can also be used to assess all transportation infrastructures’ physical condition. Additionally, non-parametric machine learning methods are applied to analyze the combined GIS, Mobile LiDAR imaging, and digital photolog big data. The results are presented to identify the most important factors affecting sign visual condition, to predict traffic sign vandalism that obstructs critical messages to drivers, and to determine factors contributing to the temporary obstruction of the sign messages. The results of data analysis provide insight to inform transportation agencies in the development of sign management plans, to identify traffic signs with a higher likelihood of failure, and to schedule sign replacement.

The image super-resolution (SR) problem is a generalization of the image restoration problem which is concerned with blur, noise, and aliasing effects. In the context of digital imaging, the purpose of image SR algorithms is to compensate for degradations such as blur resulting from camera motion and inaccurate focusing, sensor noise, and undersampling. Multiframe image SR algorithms can be used to obtain a higher-quality higher-resolution (HR) image by fusing several images that are sub-pixel-shifted versions of the same scene. By means of these algorithms, the task of super-resolving an image is often approached as an inversion problem in which a set of low-quality low-resolution (LR) images is considered to be the result of processing a high-quality high-resolution image through a dynamic image acquisition model. A special class of SR algorithms. known as 'maximum-likelihood super-resolution' (MLSR) algorithms, utilize a stochastic approach for the inversion of such a model. Basically, the HR image that is most likely to produce the observed LR images is found by solving an optimization problem. In this thesis, an overview of the most representative SR algorithms is presented. Then. the performance of two state-of-the-art MLSR algorithms based on steepest-descent optimization for grayscale and color images is evaluated and later improved by the introduction of sophisticated quasi-Newton optimization algorithms. The Davidon-Fletcher-Powell (DFP) and Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithms are appropriately reformulated to cope with the large-scale nature of the SR problem and are then applied to the SR schemes. Experimental results show that by means of the proposed algorithms, grayscale reconstruction is considerably accelerated and the quality achieved in color SR is significantly improved. Moreover, by the introduction of a practical inexact line search, the need for selecting an important parameter is eliminated. Storage-efficient variants of the BFGS algorithm are also investigated. SR algorithms based on the memoryless BFGS (MBFGS) and limited-memory BFGS (LBFGS) methods are formulated Experimental results indicate that the proposed algorithms, like the BFGS algorithm, perform the grayscale reconstruction consid¬erably faster and obtain color images of better quality. At the same time, the storage requirements for the MBFGS are comparable to those of the steepest-descent based algorithms while the LBFGS algorithm offers a meaningful trade-off between reconstructed image quality and storage requirements.

Currently, the transplantation of islets of Langerhans is a viable means to maintain control of blood sugar levels and reduce the risk of hypoglycemia in defined populations with brittle type I diabetes mellitus or those requiring pancreatectomy. However, the process of islet isolation is highly variable and not all isolations result in islet numbers or quality suitable for transplantation. This thesis aimed to improve transplantation success through optimization and standardization of the isolation process and to identify pretransplant variables associated with early islet engraftment. A previously disregarded enzyme activity, tryptic-like activity (TLA), has been identified to influence pancreas digestion efficiency and islet isolation success in both the preclinical and clinical situations. For human pancreases, islet isolation success rates improved from 0% in the lowest TLA group to over 50% in the highest TLA groups without affecting islet quality. These findings should help standardize evaluation of enzymes for clinical islet isolation. A closed, automated, pump-made gradient system was compared to the open, manual method for islet separation. No differences were observed in expected gradient volumes, islet yields or total purities between the two methods. The pump-made gradient system successfully removed manual influences on density gradient production while fulfilling regulatory requirements for closed system processing. Islet quantification was evaluated with computer-assisted digital imaging analysis (DIA) and a semi-closed assessment system. By using the DIA system method, which measures islet purity and pellet volume instead of manual counting methods, variation in islet counts and purity reduced by almost half. By using a transplant outcome measurement of C-peptide adjusted by blood glucose and creatinine, we identified four pretransplant factors that affect early transplant outcome. Of the four factors, one was related to the organ transport time, one to function of the islets, and two to the transplanted tissue volume. When these four factors were put into a predictive model, it accounted for about 40% of the transplant outcome. The work contained in this thesis identifies and optimizes a number of critical elements related to islet isolation and transplantation protocols.