This thesis examines how images are used and evaluated in a Swedish archival context. Images as of today are very common in society. They are nearly everywhere, particularly in our phones and computers. Therefore the written word is slowly losing it's sovereignty as a carrier of information. With this background the study examines if the images role also have changed in an archival setting. The following work is a qualitative research study that examines three Swedish archives: the Regional State Archive in Uppsala, Swedish Labour Movement's Archives and Library in Stockholm and the Centre for Business History in Stockholm. Furthermore the study examines how the institutions organize images and how they make them accessible for the public. It also examines what role the image play as a carrier of information in the archive. In more general terms the study also investigates what function the image can have for people in society. The comparison of the archival institutions is a methodological foundation for this study. Semi- structured interviews were conducted with archivists from the three institutions and works as the main source. For the analysis this study uses document theory and a few concepts derived from Terry Cook's archival theory, namely evidence, memory, identity and community. In terms with the previous research made on the subject this study comes to the result that images have different terms for archiving than other materials, most specifically papers documents. The examined archival institutions are today using the same tools for organizing images as they use for other documents. This can have negative effects on the ability the image has as a carrier of information. In the long run this could cause the image to become a marginalized part of the archive as a whole. The conclusion this mater's thesis presents is that the image has a potential as a source of information. Especially for groups that aren't previously represented in the archive through traditional paper documents. The image may therefore contribute to the archival pluralism and diversity, as long as the images are treated with the terms they demand as an archival document. This is a two years master's thesis in Archival Science.

Orientador: Adriano Wagner Ballarin / Coorientador: Osvaldo César Pinheiro de Almeida / Banca: Diego Augusto de Campos Moraes / Banca: Carlos Roberto Pereira Padovani / Banca: Alexandre Dal Pai / Banca: Ricardo Rall / Resumo: O setor madeireiro e toda sua cadeia produtiva possuem grande força e importância para a economia brasileira, representando 1,5% do produto interno bruto nacional em 2016. Toda madeira serrada deveria, idealmente, ser submetida a uma classificação para definição mais precisa do seu destino e justa de seu valor comercial. Quando essa madeira serrada é destinada ao exterior, a classificação é, na maioria das vezes, obrigatória. Nas serrarias do país que em sua maioria são pequenas e pouco automatizadas, a classificação é normalmente feita por visão humana, ou seja, um profissional faz a análise visual de cada peça e a classifica segundo algum critério. Como em todo processo que envolve capacidade humana, o erro é inerente e, nesse caso, elevado, em torno de 52%, segundo a literatura. Dada a importância do setor, a demanda de matéria prima e a necessidade crescente dessa classificação, é extremamente justificável que esse processo seja aperfeiçoado. A alternativa é a automatização, visando sobretudo o aumento no acerto dessa classificação. O objetivo deste trabalho foi desenvolver um modelo de redes neurais artificiais usando Deep Learning (DL) para a classificação automatizada de madeiras serradas de Pinus, seguindo as recomendações das normas da ABNT. O modelo aplicou Redes Neurais Convolucionais (Convolutional Neural Network - CNN), técnica muito estudada recentemente e promissora em diversas áreas, principalmente no processamento de imagens digitais e visão de máquina. Foram... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The timber sector and its entire production chain have great strength and importance for the Brazilian economy, representing 1.5% of the national gross domestic product in 2016. All lumber should ideally be subjected to a classification for a more precise definition of its destination and fairness of its commercial value. When this lumber is destined to the outside, classification is, in most cases, mandatory. In the country sawmills that are mostly small and little automated, the classification is usually done by human vision, that is, a professional makes the visual analysis of each piece and classifies it according to some criterion. As in any process involving human capacity, the error is inherent and, in this case, high, around 52%, according to the literature. Given the importance of the industry, the demand for raw materials and the growing need for such classification, it is extremely justifiable that this process is improved. The alternative is automation, aiming in particular to increase the accuracy of this classification. The objective of this work was to develop a model of artificial neural networks using Deep Learning (DL) for the automated classification of Pinus sawn timber, following the recommendations of ABNT standards. The model applied Convolutional Neural Network (CNN), a very recently studied and promising technique in several areas, mainly in digital image processing and machine vision. Several models were tried, being the one of better performance with accuracy of 97.50%. It was concluded that DL with CNN produces acceptable results in the classification of boards, even with few images (284), difference in the Pinus variety (elliottii and taeda) and presentation (green or dry wood, planed or not). / Doutor

Esta tese apresenta uma nova metodologia para animação de objetos líquidos em imagens. Contrariamente às técnicas existentes, este método é baseado em um modelo físico, o que proporciona efeitos realísticos. A perspectiva da imagem é obtida com a intervenção do usuário, por um esquema simples de calibração da câmera, o qual permite a projeção da camada da imagem a ser animada sobre um plano horizontal no espaço tridimensional. As equações de águas rasas conduzem a simulação e as informações de altura são projetadas de volta ao espaço da imagem utilizando traçado de raios. Além disso, efeitos de refração e iluminação são aplicados durante este estágio, resultando em animações realísticas e convincentes / This work presents a new methodology for animating liquid objects depicted in a still image. In contrast to existing techniques, the proposed method relies on a physical model to accomplish the animation, resulting in realistic effects. Image perspective is handled through a simple user assisted camera calibration scheme which allows one to project the image layers to be animated onto the horizontal plane in the three-dimensional space. Shallow-Water equations drive the simulation and the resulting height field is projected back to the image space via ray-tracing. Refraction and lighting effects are also accomplished during the ray-tracing stage, resulting in realistic and convincing animations

Úlceras cutâneas constituem um problema de saúde pública no mundo atual. A eficiência do seu tratamento é observada pela redução das áreas total, de fibrina (amarelo) e de granulação (vermelho) da úlcera, calculados manualmente e/ou por imagens, processos demorados e posteriores à consulta médica. O trabalho propõe uma nova técnica não-invasiva e automatizada de acompanhamento das úlceras por redes neurais artificiais (RNAs). Foram utilizadas imagens digitais do banco de imagens do ADUN (Ambulatório da Dermatologia de Úlceras Neurovasculares) do Hospital das Clínicas da FMRP-USP (Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo), escolhidas aleatoriamente, sendo 50 imagens para treinamento da RNA e 250 para o teste da RNA. Para validação da RNA foram criados os grupos: 1 (n=15 imagens poligonais com áreas e cores definidas previamente); 2 (n=15 imagens poligonais com áreas e cores definidas previamente, submetidas a variações de iluminação, brilho, contraste, saturação); 3 (n=15 imagens poligonais constituídas de texturas de fibrina e de granulação); 4 (n=15 imagens de úlceras cutâneas reais preenchidas totalmente em cor preta sua superfície). Para avaliar a sua aplicação clínica foram utilizadas 50 imagens padronizadas submetidas aos cálculos das áreas pela RNA. Os resultados da RNA foram comparados aos do programa Image J (segmentação manual) e/ou às medidas-padrão. Estatisticamente os programas foram considerados similares quando p > 0,05 pelo Teste t Student. Quando p < 0,05 e r positivo, considerou-se o coeficiente de correlação de Pearson. A base de imagens de úlceras cutâneas foi eficiente para a aquisição das imagens, para a criação e execução dos algoritmos de extração de cores, de treinamento e de teste da RNA. A rede neural artificial desenvolvida apresentou desempenho similar ao Image J e às medidas-padrão adotadas para a segmentação das figuras do grupo 1, sendo p > 0,05 para as áreas total, de fibrina e de granulação. Na avaliação de interferência de ruídos (grupo 2), foi verificado que tais fatores não interferiram na segmentação da área dos polígonos (p > 0,05), pela RNA e pelo Image J. Entretanto, apesar de interferirem na segmentação de cores de granulação, sendo p < 0,05, o coeficiente de correlação RNA/Image J foi de 0,90 com p < 0,0001. No grupo 3, os cálculos das áreas foram semelhantes pela RNA e pelo Image J (p > 0,05). Quando comparadas às áreas calculadas pelos programas às medidas-padrão, o coeficiente de correlação foi significante (p < 0,0001) para todas as áreas. A segmentação das áreas das úlceras do grupo 4 pela RNA foi validada quando comparada à segmentação manual pelo Image J (p> 0,05). A aplicação clínica da RNA sobre o banco de imagens foi semelhante ao Image J para a segmentação das áreas (p > 0,05). Enfim, a rede neural artificial desenvolvida no Matlab 7.0 mostrou desempenho eficaz e validado na segmentação das úlceras de perna quanto à automatização do cálculo das áreas total, de fibrina e de granulação, semelhante à oferecida manualmente pelo programa Image J. Além disso, mostrou-se de grande aplicação clínica devido a facilidade de sua utilização através da interface web criada, sua praticidade, não interferência do usuário (automatização), propriedades essas que a consolida como uma metodologia adequada para o acompanhamento dinâmico-terapêutico da evolução das úlceras cutâneas. / Cutaneous ulcers are a public health problem worldwide. The efficiency of their treatment is observed through the reduction on the total affected areas, slough (yellow) and granulation (red) of the ulcer, manually calculated and/or through images, which are delayed processes usually performed after medical consultation. This work proposes a new non-invasive and automated technique to follow-up ulcers through artificial neural networks (ANN). Digital images from the ADUN (Neurovascular Ulcers Dermatology Ambulatory) image bank - FMRP General Hospital (Ribeirão Preto Medical School - University of São Paulo) were used and randomly selected as follows: 50 images for ANN training and 250 for the ANN test. For the ANN validation, the following groups were created: 1 (n=15 polygonal images with areas and colors previously defined); 2 (n=15 polygonal images with areas and colors previously defined submitted to illumination, brightness, contrast and saturation variation); 3 (n=15 polygonal images composed of slough and granulation textures); 4 (n=15 images of actual cutaneous ulcers with their surface fully filled in black). To evaluate its clinical application, 50 standard images were used and submitted to calculation of areas using ANN. The ANN results were compared to those obtained with the Image J software (manual segmentation) and/or to standard measures. The programs were statistically considered similar when p > 0.05 through the t Student test. When p < 0.05 and r is positive, the Pearson correlation coefficient was considered. The cutaneous ulcer image bank was efficient for the acquisition of images, for the creation and execution of color extraction algorithms, ANN training and tests. The artificial neural network developed presented performance similar to that obtained with the Image J software and to standard measures adopted for the segmentation of figures from group 1, with p > 0.05 for total areas, slough and granulation. In the noise interference assessment (group 2), it was verified that such factors did not interfere in the polygons area segmentation (p > 0.05) through both ANN and Image J. However, although interfering in the color and granulation segmentation, with p < 0.05, the ANN/Image J correlation coefficient was of 0.90, with p < 0.0001. In group 3, the calculations of areas were similar through both ANN and Image J (p > 0.05). When compared to standard measures, the correlation coefficient was significant (p < 0.0001) for all areas. The segmentation of ulcer areas of group 4 through ANN was validated when compared to manual segmentation through Image J (p> 0.05). The clinical application of ANN on the image bank was similar to Image J for the segmentation of areas (p > 0.05). Finally, the Artificial Neural Network developed in Matlab 7.0 environment showed good performance and was validated in the segmentation of leg ulcers in relation to the automation of the calculation of total areas, slough and granulation, which was similar to that obtained with the Image J software. Moreover, it presented a large clinical application due to the easiness of its application through the web interface created and the non interference of the user (automation), properties that consolidate this technique as a suitable methodology for the dynamic-therapeutic follow-up of the evolution of cutaneous ulcers.

Úlceras cutâneas constituem um problema de saúde pública no mundo atual. A eficiência do seu tratamento é observada pela redução das áreas total, de fibrina (amarelo) e de granulação (vermelho) da úlcera, calculados manualmente e/ou por imagens, processos demorados e posteriores à consulta médica. O trabalho propõe uma nova técnica não-invasiva e automatizada de acompanhamento das úlceras por redes neurais artificiais (RNAs). Foram utilizadas imagens digitais do banco de imagens do ADUN (Ambulatório da Dermatologia de Úlceras Neurovasculares) do Hospital das Clínicas da FMRP-USP (Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo), escolhidas aleatoriamente, sendo 50 imagens para treinamento da RNA e 250 para o teste da RNA. Para validação da RNA foram criados os grupos: 1 (n=15 imagens poligonais com áreas e cores definidas previamente); 2 (n=15 imagens poligonais com áreas e cores definidas previamente, submetidas a variações de iluminação, brilho, contraste, saturação); 3 (n=15 imagens poligonais constituídas de texturas de fibrina e de granulação); 4 (n=15 imagens de úlceras cutâneas reais preenchidas totalmente em cor preta sua superfície). Para avaliar a sua aplicação clínica foram utilizadas 50 imagens padronizadas submetidas aos cálculos das áreas pela RNA. Os resultados da RNA foram comparados aos do programa Image J (segmentação manual) e/ou às medidas-padrão. Estatisticamente os programas foram considerados similares quando p > 0,05 pelo Teste t Student. Quando p < 0,05 e r positivo, considerou-se o coeficiente de correlação de Pearson. A base de imagens de úlceras cutâneas foi eficiente para a aquisição das imagens, para a criação e execução dos algoritmos de extração de cores, de treinamento e de teste da RNA. A rede neural artificial desenvolvida apresentou desempenho similar ao Image J e às medidas-padrão adotadas para a segmentação das figuras do grupo 1, sendo p > 0,05 para as áreas total, de fibrina e de granulação. Na avaliação de interferência de ruídos (grupo 2), foi verificado que tais fatores não interferiram na segmentação da área dos polígonos (p > 0,05), pela RNA e pelo Image J. Entretanto, apesar de interferirem na segmentação de cores de granulação, sendo p < 0,05, o coeficiente de correlação RNA/Image J foi de 0,90 com p < 0,0001. No grupo 3, os cálculos das áreas foram semelhantes pela RNA e pelo Image J (p > 0,05). Quando comparadas às áreas calculadas pelos programas às medidas-padrão, o coeficiente de correlação foi significante (p < 0,0001) para todas as áreas. A segmentação das áreas das úlceras do grupo 4 pela RNA foi validada quando comparada à segmentação manual pelo Image J (p> 0,05). A aplicação clínica da RNA sobre o banco de imagens foi semelhante ao Image J para a segmentação das áreas (p > 0,05). Enfim, a rede neural artificial desenvolvida no Matlab 7.0 mostrou desempenho eficaz e validado na segmentação das úlceras de perna quanto à automatização do cálculo das áreas total, de fibrina e de granulação, semelhante à oferecida manualmente pelo programa Image J. Além disso, mostrou-se de grande aplicação clínica devido a facilidade de sua utilização através da interface web criada, sua praticidade, não interferência do usuário (automatização), propriedades essas que a consolida como uma metodologia adequada para o acompanhamento dinâmico-terapêutico da evolução das úlceras cutâneas. / Cutaneous ulcers are a public health problem worldwide. The efficiency of their treatment is observed through the reduction on the total affected areas, slough (yellow) and granulation (red) of the ulcer, manually calculated and/or through images, which are delayed processes usually performed after medical consultation. This work proposes a new non-invasive and automated technique to follow-up ulcers through artificial neural networks (ANN). Digital images from the ADUN (Neurovascular Ulcers Dermatology Ambulatory) image bank - FMRP General Hospital (Ribeirão Preto Medical School - University of São Paulo) were used and randomly selected as follows: 50 images for ANN training and 250 for the ANN test. For the ANN validation, the following groups were created: 1 (n=15 polygonal images with areas and colors previously defined); 2 (n=15 polygonal images with areas and colors previously defined submitted to illumination, brightness, contrast and saturation variation); 3 (n=15 polygonal images composed of slough and granulation textures); 4 (n=15 images of actual cutaneous ulcers with their surface fully filled in black). To evaluate its clinical application, 50 standard images were used and submitted to calculation of areas using ANN. The ANN results were compared to those obtained with the Image J software (manual segmentation) and/or to standard measures. The programs were statistically considered similar when p > 0.05 through the t Student test. When p < 0.05 and r is positive, the Pearson correlation coefficient was considered. The cutaneous ulcer image bank was efficient for the acquisition of images, for the creation and execution of color extraction algorithms, ANN training and tests. The artificial neural network developed presented performance similar to that obtained with the Image J software and to standard measures adopted for the segmentation of figures from group 1, with p > 0.05 for total areas, slough and granulation. In the noise interference assessment (group 2), it was verified that such factors did not interfere in the polygons area segmentation (p > 0.05) through both ANN and Image J. However, although interfering in the color and granulation segmentation, with p < 0.05, the ANN/Image J correlation coefficient was of 0.90, with p < 0.0001. In group 3, the calculations of areas were similar through both ANN and Image J (p > 0.05). When compared to standard measures, the correlation coefficient was significant (p < 0.0001) for all areas. The segmentation of ulcer areas of group 4 through ANN was validated when compared to manual segmentation through Image J (p> 0.05). The clinical application of ANN on the image bank was similar to Image J for the segmentation of areas (p > 0.05). Finally, the Artificial Neural Network developed in Matlab 7.0 environment showed good performance and was validated in the segmentation of leg ulcers in relation to the automation of the calculation of total areas, slough and granulation, which was similar to that obtained with the Image J software. Moreover, it presented a large clinical application due to the easiness of its application through the web interface created and the non interference of the user (automation), properties that consolidate this technique as a suitable methodology for the dynamic-therapeutic follow-up of the evolution of cutaneous ulcers.

The field of telehealth has developed rapidly in recent years. It provides medical support particularly to those who are living in remote areas and in emergency cases. Although developments in both technology and practice have been rapid, there are still many gaps in our knowledge with regard to the effective application of telehealth. This study investigated human colour perception in telehealth, specifically the colour red as one of the key symptoms when diagnosing different pathologies. The quality of medical images is safety critical when transmitting the symptoms of pathologies in telehealth, as distorted or degraded colours may result in errors. The study focused on the use of digital images in teleconsultation, particularly on images showing cellulitis (bacterial skin infection) and conjunctivitis (red eye) as case studies, as both of these pathologies involve the colour red in their diagnosis. The study proposed and tested the use of an image quality scale, which represented the level of image resolution; a red colour scale, which represented the intensity of redness in an image; and a confidence scale, which represented the levels of confidence that telehealth users had when judging the colour red. The research involved a series of experiments using hypothetico-deductive and formal hypothesis testing with two groups of participants, medical doctors and non-medical participants. The experiments were conducted in collaboration with the local National Health Service (NHS) Accident and Emergency (A&E) department at Aberdeen Royal Infirmary (ARI). Medical experts in ophthalmology and dermatology were also involved in selecting and verifying the relevant images. The study found that doctors and non-doctors were consistent in the majority of the experiments. The accuracy of the participants was demonstrably higher when using a colour scale with pictures, more so for the non-doctor group than the doctor group. It also found that the level of accuracy for both doctors and nondoctors was higher when using red colour scale of three divisions than when using a scale of five divisions. This result was supported by previous studies, which used telehealth for diagnosing extreme cases. The study also found that when the image quality was poor the participants had higher error rates and less consistency in their answers. The study found poor correlation between accuracy, confidence and time for both participant groups. The study found that most participants in both doctor and non-doctor groups had high confidence most of the time, whether the accuracy was high or low. It was also found that medical background or clinical experience had no effect on the accuracy level across the experiment sets. In some cases, doctors with no or little experience had higher accuracy than those with greater experience. This result may have significant implications for the feasibility of involving non-doctors in the management of telehealth systems, especially in tasks not requiring medical skills, such as colour classification. This has the potential to provide a considerable saving in resources and costs for healthcare providers. An auto-evaluation system was introduced, and proposed for further study, in order to improve the current telehealth diagnostic protocol and to avoid or prevent errors by making red colour classification more objective and accurate.

Jiang, Lei. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (p. 68-75). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Background and Motivation --- p.5 / Chapter 2.1 --- Foreshortening --- p.5 / Chapter 2.1.1 --- Vanishing Point --- p.6 / Chapter 2.1.2 --- Metric Rectification --- p.8 / Chapter 2.2 --- Content Aware Image Resizing --- p.11 / Chapter 2.3 --- Texture Deformation --- p.15 / Chapter 2.3.1 --- Shape from texture --- p.16 / Chapter 2.3.2 --- Shape from lattice --- p.18 / Chapter 3 --- Resizing on Facade --- p.21 / Chapter 3.1 --- Introduction --- p.21 / Chapter 3.2 --- Related Work --- p.23 / Chapter 3.3 --- Algorithm --- p.24 / Chapter 3.3.1 --- Facade Detection --- p.25 / Chapter 3.3.2 --- Facade Resizing --- p.32 / Chapter 3.4 --- Results --- p.34 / Chapter 4 --- Cell Texture Editing --- p.42 / Chapter 4.1 --- Introduction --- p.42 / Chapter 4.2 --- Related Work --- p.44 / Chapter 4.3 --- Our Approach --- p.46 / Chapter 4.3.1 --- Cell Detection --- p.47 / Chapter 4.3.2 --- Local Affine Estimation --- p.49 / Chapter 4.3.3 --- Affine Transformation Field --- p.52 / Chapter 4.4 --- Photo Editing Applications --- p.55 / Chapter 4.5 --- Discussion --- p.58 / Chapter 5 --- Conclusion --- p.65 / Bibliography --- p.67

Made available in DSpace on 2016-03-15T19:43:05Z (GMT). No. of bitstreams: 1 Fernanda Maria Oliveira Araujo.pdf: 52133006 bytes, checksum: e885f484f4306b9d7538e1ec6fc2c3f7 (MD5) Previous issue date: 2015-06-22 / This dissertation proposes an analysis on the digitalization of visual artworks collections, understood as collections of traditional art s objects such as paintings, sculptures, prints and drawings starting from the decision to reproduce this artwork in a digital object up to the final view on virtual environment as the internet. During the studies it is observed that a collection digitalization, regardless of its nature, is more than an ordinary digital conversion, requiring an structured process that impacts systemically, functionally and organizationally all levels of the institution that decides to adhere to digital technologies. Therefore, the artworks collections digitalization goes beyond the issues inherent to scan text documents widespread in libraries. The visual objects (artworks) with its aesthetic contextualization must be subject to reproduction and digitalization processes that consider its intrinsic elements and oriented interfaces to promote users experience. Eventually, the transposition of the museum institutions responsibility as cultural heritage curators of society to a relevant channel for the democratization of information and knowledge diffusion proves to be a reality, a new path with no return. / Esta dissertação propõe uma reflexão sobre a digitalização de acervos de obras de artes visuais, entendidos como coleções de objetos tradicionais da arte tais como pinturas, esculturas, gravuras e desenhos. A trajetória de digitalização desde a decisão da reprodução da obra de arte em um objeto digital à visualização dele em um ambiente virtual como a internet. Ao longo dos estudos observa-se que a digitalização de um acervo, independente de sua natureza, é mais que um processo técnico de conversão digital pois representa um processo estruturado e estratégico que impacta sistemicamente, funcionalmente e organizacionalmente a instituição que decide aderir às tecnologias digitais, ou seja, a digitalização de obras de artes visuais extrapola as questões inerentes à digitalização de documentos textuais praticados em bibliotecas, os objetos visuais (as obras de arte) com seus potenciais estéticos, devem ser pensados a partir de processos de reprodução e de visualização que considerem estes elementos, como através de interfaces esteticamente elaboradas que envolvam os usuários. Constata-se, ao final, que a transposição do papel da instituição museológica de cuidadora do patrimônio cultural da sociedade na qual está inserida para o de canal de democratização da informação e difusão do conhecimento mostra-se uma realidade, um caminho sem volta.

Deep generative modeling is an increasingly popular area of machine learning that takes advantage of recent developments in neural networks in order to estimate the distribution of observed data. In this dissertation we introduce three advances in this area. The first one, Maximum Entropy Flow Networks, allows to do maximum entropy modeling by combining normalizing flows with the augmented Lagrangian optimization method. The second one is the continuous Bernoulli, a new [0,1]-supported distribution which we introduce with the motivation of fixing the pervasive error in variational autoencoders of using a Bernoulli likelihood for non-binary data. The last one, Deep Random Splines, is a novel distribution over functions, where samples are obtained by sampling Gaussian noise and transforming it through a neural network to obtain the parameters of a spline. We apply these to model texture images, natural images and neural population data, respectively; and observe significant improvements over current state of the art alternatives.

This thesis aims to examine the issues involved in utilising digital images and assess whether image processing techniques can be used as a cost-effective method of reconstructing the image found in a deteriorated cellulose acetate negative. Negatives affected by the vinegar syndrome are found in large numbers within Australian institutions. This was confirmed by a survey (using a questionnaire) undertaken at the National Library of Australia in 2000. The survey also found that although these collections are large, and hence the level of deterioration variable, little could be done to restore any of the negatives once deterioration had begun. Storing negatives at low temperature and low relative humidity slows down the breakdown of cellulose acetate; however, it cannot reverse the process once it has commenced. Although removing the gelatine pellicular from the deteriorated cellulose acetate support (making the image easier to view) a possible method of restoration, this becomes unfeasible when dealing with a large collection. As a result, how to manage cellulose acetate negatives once they have deteriorated becomes problematic. Image-processing techniques used to digitally restore these negatives were examined via a series of case studies. These examinations were undertaken using two software packages-the Image Processing Tool kit (IPTK) and OPTIMAS. Deteriorated cellulose acetate negatives were scanned, then a number of program filters were applied to the digital image to determine whether disfiguring elements (referred to as channelling elements) resulting from the deteriorated support could be digitally removed. IPTK and OPTIMAS were not completely successful in removing the deteriorated elements from the digital version. The results highlighted that a number of issues relating to the use of digital technology needed to be addressed. These issues included knowledge of basic technical terms, an understanding of digital language, and how to include the use of digital technology into a long-term strategy for archiving a digitised collection. This thesis showed that issues relating to utilising digital systems could be addressed by implementing a preservation management plan. A preservation management plan can be used to incorporate the goals of digitising, the long-term issues of retaining digital files, ongoing access relating to the digital file, hardware and software, and the importance of having the relevant expertise when undertaking such a project. Due to the limitations of the printed hardcopy displaying features in a number of the images (figures) outlined in this thesis, a compact disk (CD) has been included with this submission and can be found at the end of this document.