La presente investigación propone un dispositivo electrónico portátil que permite el diagnóstico rápido de enfermedades utilizando tiras rápidas o ensayos de flujo lateral (LFAs), el cual indica si el resultado es positivo, negativo o inválido, dichas tiras utilizan como muestra el suero de sangre. Actualmente existen soluciones similares con un costo muy elevado, por lo que en el Laboratorio de Biología Molecular del Instituto de Investigación de Enfermedades Infecciosas (IIEI) se realiza el diagnóstico por inspección visual; como consecuencia de la fatiga visual, se reduce la confiabilidad de los resultados proporcionados por los técnicos del laboratorio. En la literatura científica se propone como solución un equipo que requiere de una cámara de celular, iluminación con láser, el uso de algoritmos de visión artificial y de redes neuronales convolucionales. El presente equipo utiliza algoritmos computacionales de procesamiento digital de imágenes sobre una foto tomada por la cámara digital integrada, recorte del área de interés, construcción de histogramas y adquisición de descriptores de los componentes R, G y B de la imagen, que serán utilizados por una red neuronal multicapa. Dicho software se complementa con un hardware que incluye un computador de placa reducida, el cual emitirá el resultado a través de una pantalla táctil e interactiva. De acuerdo con lo obtenido al emplear el algoritmo, se consiguió un grado de acuerdo superior a 0.9, para esto se utilizó el índice de Kappa Cohen, coincidiendo mayormente los resultados emitidos por los técnicos de laboratorio con los obtenidos por la solución propuesta. / This dissertation proposes a portable electronic device that allows the rapid diagnosis of diseases using rapid immunochromatographic test or lateral flow assays (LFAs) which indicates whether the result is positive, negative, or invalid, these strips use blood serum as a sample. Nowadays there are similar solutions to which they have a very high cost, so the diagnosis is made by visual inspection at the Laboratorio de Biología Molecular del Instituto de Investigación de Enfermedades Infecciosas (IIEI); because of visual fatigue, the reliability of the results provided by laboratory technicians is reduced. In the scientific literature, the design of a device that requires a cell phone camera, laser lighting, the use of artificial vision algorithms and convolutional neural networks is proposed as a solution. This equipment uses computational algorithms for digital image processing on a photo taken by the integrated digital camera, such as trimming the area of ​​interest, constructing histograms and acquisition of descriptors of the R, G and B components of the image taken by the camera, which will be used by a multilayer neural network. This software is complemented by hardware that includes a reduced board computer, which will issue the results through an interactive touch screen. According to what was obtained when using the algorithm, a degree of agreement greater than 0.9 was achieved, for this the Kappa Cohen index was used, mostly coinciding with the results issued by the laboratory technicians with those obtained by the proposed solution. / Tesis

The thesis consists of three parts. Theoretical description of digital image processing, optical character recognition and design of system for car licence plate recognition (LPR) in image or video sequence. Theoretical part describes image representation, smoothing, methods used for blob segmentation and proposed are two methods for optical character recognition (OCR). Concern of practical part is to find solution and design procedure for LPR system included OCR. The design contain image pre-processing, blob segmentation, object detection based on its properties and OCR. Proposed solution use grayscale trasformation, histogram processing, thresholding, connected component,region recognition based on its patern and properties. Implemented is also optical recognition method of licence plate where acquired values are compared with database used to manage entry of vehicles into object.

Detta projekt åsyftade att möjliggöra för en industrirobot att illustrera godtyckliga digitalabilder på en plan yta. Detta uppnåddes genom att utrusta en manipulator med ett ritverktyg. Genom digital bildbehandling så kunde rörelsemönster genereras vilka matades till industriroboten för att den skulle kunna återskapa den digitala versionen. Roboten ritar med en teknik benämnd pointillism som innebär att endast punkter plottas. Resultatet blir en konkret svartvit representation av originalbilden. Projektet genomfördes på institutionen Industriell Produktion på Kungliga Tekniska Högskolan. Projektet är i sin natur väldigt inriktat på forskning och utveckling eftersom det går ut på skapandet av en teknik för att uppnå ett tydligt mål. Kontinuerlig utveckling var kopplat till målet för att förbättra resultatet från olika aspekter. / This project strived to enable an industrial robot to illustrate arbitrary digitized images on a planar surface. This was accomplished by equipping a robotic manipulator with a drawing utensil. Motion patterns were generated based on digital image processing and fed to the robot for it to imitate the digital version. The robot prints with a technique called pointillism, which implies that solely points are plotted. The result is a tangible black and white representation of the original image. The project was carried out in the Production Engineering facilities at the Royal Institute of Technology. The nature of the project is very research and development oriented as it deals with the creation of a technology to achieve an explicit goal. Continuous development was related to the goal to improve the result from different aspects.

Concluypo enmbargo. / Solicitud de envío manuscrito de artículo científico. / El presente trabajo propone un dispositivo electrónico portátil que, acoplado a un microscopio óptico convencional, pueda capturar una imagen microscópica de la fibra óptica y emita un juicio si el conector se encuentra en un estado de limpieza aceptable o no. Asimismo, mediante reportes de trabajo y usando normas internacionales, evaluar los conectores para que formen parte de una base de datos histórica que aporte a los mantenimientos preventivos de una red. Si bien en la actualidad se tienen soluciones tecnológicas parecidas, estas tienen como gran desventaja el alto costo que implica adquirir uno de estos equipos, desanimando a los operadores optar por este importante equipo. Ingenierilmente, esta propuesta será posible mediante el desarrollo de un algoritmo que procese las imágenes capturadas con el uso de diversas herramientas de procesamiento de imágenes como la binarización, transformada de Hough, algoritmo de Canny, escala de grises y ecualización de imágenes, que, montado sobre un hardware basado en un computador de placa reducida, entregará una evaluación del conector de fibra óptica y un reporte descargable que sirve como documentación del trabajo realizado. Dentro de este contexto, los resultados obtenidos son favorables para una utilización en campo, para llegar a esto se tuvo como como variable de calidad el índice de Kappa Cohen, en el cual se demuestra la alta tasa de coincidencia entre las evaluaciones del equipo y la evaluación de ingenieros expertos. / The present work proposes a portable electronic device that, coupled with a conventional optical microscope, can capture a microscopic image of the optical fiber and emit a judgment if the connector is in an optimal cleaning state or not. Likewise, through reports and using international standards, evaluate the connectors so that can be part of a source of database that contributes to the preventive maintenance of a network. Although similar technological solutions are currently available, these have the great disadvantage of the high cost of acquiring one of this equipment, discouraging operators from opting for this important procedure. Engineeringfully, this proposal will be possible through the development of an algorithm that processes the captured images with the use of various image processing tools such as binarization, Hough transform, Canny algorithm, grayscale and image equalization, which, mounted on a hardware based on a reduced plate computer, it will provide an evaluation of the fiber optic connector and a downloadable report that serves as documentation of the work done. Within this context, the results obtained are favorable for a field use, to achieve this, the Kappa Cohen index was used as a quality variable, which demonstrates the high rate of coincidence between the evaluations of the equipment and the Evaluation of expert engineers. / Tesis

The thesis is aimed on the digital analysis and processing of micro- scopic image data with a focus on cross-section samples from the artworks which fall into cultural heritage domain. It contributes to solution of two different problems of image processing - image seg- mentation and image retrieval. The performance evaluation of differ- ent image segmentation methods on a data set of cross-section images is carried out in order to study the behavior of individual approaches and to propose guidelines how to choose suitable method for segmen- tation of microscopic images. Moreover, the benefit of segmenta- tion combination approach is studied and several distinct combination schemes are proposed. The evaluation is backed up by a large number of experiments where image segmentation algorithms are assessed by several segmentation quality measures. Applicability of achieved re- sults is shown on image data of different origin. In the second part, content-based image retrieval of cross-section samples is addressed and functional solution is presented. Its implementation is included in Nephele system, an expert system for processing and archiving the material research reports with image processing features, designed and implemented for the cultural heritage application area. 1

El presente trabajo de tesis propone el diseño de un sistema de diagnóstico de la calidad de piezas de melamina que mediante el procesamiento de imágenes y la aplicación de técnicas y herramientas de aprendizaje profundo mejorará considerablemente el proceso de control de calidad de una empresa procesadora de melamina. Para lograr el resultado esperado se realizaron un conjunto de actividades que incluyen principalmente la planificación del proyecto, el análisis del negocio y la elaboración de la propuesta de solución. La planificación del proyecto toma como base las prácticas incluidas en la Guía PMBOK®, 6ta edición; e incluye la elaboración de una serie de planes que permitieron gestionar distintos aspectos del proyecto como el alcance, cronograma, costos, riesgos, comunicaciones, entre otros. Por otro lado, para realizar el análisis de la situación actual de la empresa, se utilizó como referencia el marco de trabajo de Zachman, el cual permite analizar los procesos de negocio de la organización desde diferentes perspectivas e identificar los obstáculos que cada uno de los procesos enfrenta y que se buscan corregir con la presente propuesta de solución. Finalmente, para elaborar la propuesta de solución, se identificaron y priorizaron los drivers funcionales, no funcionales y de restricción. Las decisiones de diseño se toman sobre estos drivers y se documentan usando los diagramas del modelo C4. / This thesis work proposes the design of a diagnosis system for the quality of melamine parts that through image processing and the application of deep learning techniques and tools will considerably improve the quality control process of a melamine manufacturing company. To achieve the expected result, a set of activities was performed including primarily project planning, business analysis and development of the solution proposal. Project planning is based on the practices from PMBOK® Guide, 6th edition, and includes the development of a series of plans that allows the management of different aspects of the project such as scope, schedule, costs, risks, communications, among others. On the other hand, to analyze the company's current situation, the Zachman framework was used as a reference allowing for analysis of the organization's business processes from different perspectives and identification of the obstacles that each of the processes face and those that are sought to be corrected with this proposed solution. Finally, to prepare the proposed solution, functional, non-functional and restrictions drivers were identified and prioritized. Design decisions are made on these drivers and documented using the C4 model diagrams. / Tesis

With the growth of computer power, Digital Image Processing plays a more and more important role in the modern world, including the field of industry, medical, communications, spaceflight technology etc. There is no clear definition how to divide the digital image processing, but normally, digital image processing includes three main steps: low-level, mid-level and highlevel processing. Low-level processing involves primitive operations, such as: image preprocessing to reduce the noise, contrast enhancement, and image sharpening. Mid-level processing on images involves tasks such as segmentation (partitioning an image into regions or objects), description of those objects to reduce them to a form suitable for computer processing, and classification (recognition) of individual objects. Finally, higher-level processing involves "making sense" of an ensemble of recognised objects, as in image analysis. Based on the theory just described in the last paragraph, this thesis is organised in three parts: Colour Edge and Face Detection; Hand motion detection; Hand Gesture Detection and Medical Image Processing. II In Colour Edge Detection, two new images G-image and R-image are built through colour space transform, after that, the two edges extracted from G-image and R-image respectively are combined to obtain the final new edge. In Face Detection, a skin model is built first, then the boundary condition of this skin model can be extracted to cover almost all of the skin pixels. After skin detection, the knowledge about size, size ratio, locations of ears and mouth is used to recognise the face in the skin regions. In Hand Motion Detection, frame differe is compared with an automatically chosen threshold in order to identify the moving object. For some special situations, with slow or smooth object motion, the background modelling and frame differencing are combined in order to improve the performance. In Hand Gesture Recognition, 3 features of every testing image are input to Gaussian Mixture Model (GMM), and then the Expectation Maximization algorithm (EM)is used to compare the GMM from testing images and GMM from training images in order to classify the results. In Medical Image Processing (mammograms), the Artificial Neural Network (ANN) and clustering rule are applied to choose the feature. Two classifier, ANN and Support Vector Machine (SVM), have been applied to classify the results, in this processing, the balance learning theory and optimized decision has been developed are applied to improve the performance.

O crescimento da utilização de sistemas de reconhecimento no mundo contemporâneo exige processos de detecção cada vez mais robustos e ágeis. Aplicáveis desde sistemas de teleconferência empresarial até mecanismos de segurança e vigilância, a detecção e o reconhecimento de pessoas tornaram-se uma constante. Na tentativa de buscar caminhos alternativos, tanto para os problemas de detecção, quanto para os de reconhecimento, este trabalho propõe a utilização de medidas biométricas, mensuradas em imagens digitalizadas de faces humanas. A partir do estudo de tais medidas, torna-se possível a verificação de proporções existentes na face, especialmente a proporção divina, podendo constituir, no futuro, a base para algoritmos de detecção e/ou reconhecimento que usufruam das informações trazidas por tais proporções. Diante de uma reduzida quantidade de publicações no meio científico que utilizam a proporção divina como meio de detecção e/ou reconhecimento em processamento de imagens, esta investigação vem contribuir com alguns passos nessa direção / The increase of recognition systems in the contemporary world has demanded robust and agile detection processes. From teleconference systems to security and monitoring mechanisms, the detection and recognition of people have became constantly used and applied. In attempt to search for alternative ways to solve both detection and recognition problems, this work proposes the utilization of biometric measures, taken in digital image of human faces. From the study of such measures, it’s possible to verify face proportions, especially the divine proportion, which could allows, in the future, to implement the detection and/or recognition algorithms that utilize such proportions. Due to small amount of scientific publications that use the divine proportion as a way of detection and/or recognition in image processing, this investigation contributes with some steps in this direction

Nesta dissertação é abordado o tema da avaliação de qualidade em sinais de vídeo, especificamente da avaliação objetiva completamente referenciada de sinais de vídeo em definição padrão. A forma mais confiável de se medir a diferença de qualidade entre duas cenas de vídeo é utilizando um painel formado por telespectadores, resultando em uma medida subjetiva da diferença de qualidade. Esta metodologia demanda um longo período de tempo e um elevado custo operacional, o que a torna pouco prática para utilização. Neste trabalho são apresentados os aspectos relevantes do sistema visual humano, das metodologias para avaliação de vídeo em aplicações de televisão digital em definição padrão e também da validação destas metodologias. O objetivo desta dissertação é testar métricas de baixo custo computacional como a que avalia a relação sinal-ruído de pico (PSNR: Peak Signal-to-Noise Ratio), a que mede similaridade estrutural (SSIM: Structural SIMilarity) e a que mede diferenças em três componentes de cor definidas pela CIE (Commission Internationale de l\'Eclairage), representadas por L*, a* e b* em uma dada extensão espacial (S-CIELAB: Spatial-CIELAB). Uma metodologia de validação destas métricas é apresentada, tendo como base as cenas e resultados dos testes subjetivos efetuados pelo Grupo de Especialistas em Qualidade de Vídeo (VQEG: Video Quality Expert Group). A estas métricas é introduzida uma etapa de preparação das cenas, na qual são efetuadas equalização de brilho, suavização de detalhes e detecção de contornos. Controlando-se a intensidade destes filtros, um novo conjunto de medidas é obtido. Comparações de desempenho são realizadas entre estes novos conjuntos de medidas e o conjunto de medidas obtido pelo VQEG. Os resultados mostram que para aplicações em televisão digital de definição padrão, a avaliação utilizando componentes de cor pouco influencia na correlação com as medidas obtidas nos testes subjetivos. Por outro lado, foi verificado que a aplicação adequada de técnicas para suavização de imagens, combinadas com métricas de fácil implementação como a SSIM, elevam seu grau de correlação com medidas subjetivas. Também foi demonstrado que técnicas para extração de contornos, combinadas com a métrica PSNR, podem aumentar significativamente seu desempenho em termos de correlação com os testes efetuados pelo VQEG. À luz destes resultados, foi concluído que medidas objetivas de fácil implementação do ponto de vista computacional podem ser usadas para comparação da qualidade de sinais de vídeo SDTV, desde que devidamente combinadas com técnicas para adequação ao sistema visual humano como a suavização e extração de contornos. / This research is about the video signal quality comparison issue, focusing at full reference metrics using standard definition television. The most reliable way to predict the differences in terms of quality between two video scenes is using a panel of television viewers, under controlled psychometric experimental conditions, resulting in statistical meaningful Differences in Mean Opinion Score (DMOS). The Subjective assessment is both time consuming and costly, therefore with practical limitations. The ideal substitute are objective quality assessment algorithms, whose scores have been shown to correlate highly with the results of DMOS. The goal for this research is to optimize the performance of simple metrics combining it with digital image processing. First this work presents many relevant aspects of the human visual system, methodologies for video evaluation in digital television applications using standard definition (SDTV) and also a validation methodology of these methods. After that, the main goal is to test three very simple metrics in terms of computational cost: PSNR (Peak Signal-to-Noise Ratio), SSIM (Structural SIMilarity) and S-CIELAB (Spatial-CIELAB). original metrics were modified in order to improve their correlations against subjective assessment data. Several experiments combining the advantages of digital image filters for softness and edge extraction have been accomplished within this work. The results show that such simple metrics combined with digital image processing for edge extraction, for example, do improve their correlations with subjective assessment.

Made available in DSpace on 2017-07-21T14:19:29Z (GMT). No. of bitstreams: 1 TIAGO GERKE.pdf: 2494737 bytes, checksum: 62b5d2787dcdebe0e9ccf39bfa75f626 (MD5) Previous issue date: 2017-02-03 / Wheat cultivation plays an important role to Brazil and the world economic development, as well as in the human diet. The wheat Brazilian production is insufficient to meet the national demand, making research needed in order to improve the yield of this cereal. The goal of this work was to estimate wheat yield, searching for a predictive model through the data mining techniques, with data obtained from high spatial resolution images collected by unmanned aerial vehicles (UAV). The work was carried out in two experimental areas at Ponta Grossa city, Parana state, where for each area eight images were taken, at different culture development stages, with spatial resolution of 3.4cm/px and two images with resolution 10cm/px and 20cm/px, using an eBee UAV with an RGB and a NIR camera. The image processing was done with the Pix4D software, and resulted in an orthomosaics with reflectance values at different wavelengths: Red, Green and Blue, from the RGB camera and Red, Greed and NIR from the NIR camera, besides an image with NDVI values obtained from the arithmetic of NIR and Red wavelengths. The georeferencing correction of each orthomosaic and the extraction of the reflectance values were done with Quantum GIS geographic information system (GIS). From the extracted reflectance values, databases in different proportions (10%, 20%, 40%, 70% and 100%) were created for data mining, using the SMOReg algorithm, based on a support vector machine (SVM) for regression (SVR). The georreferencing correction using 10 control points provided ortomosaics with mean square error (RSME) of distance of 0.35m, which did not show significant difference compared to the correction with 5 control points (RMSE = 0.38m). The reflectance values were different for each study area, making it difficult to indicate better periods for estimating wheat yield. The highest correlation were obtained with data from RGB camera images, followed by the NIR and NDVI camera, with correlations of 0.6168,0.5423 and 0.5324, respectively. The amount of information extracted from the images, reflected in the proportion of the databases, was not significant to generated predictive models, as well as in the correlation indexes, which were statistically the same. Better correlation indices were obtained from the data extracted from the images with spatial resolution of 20cm/px, which suggests that high spatial resolution images may not be adequate for wheat yield estimation. / O cultivo do trigo desempenha um papel importante no desenvolvimento econômico de várias regiões do Brasil e do mundo, bem como na dieta humana. A produção brasileira do trigo é insuficiente para atender à demanda nacional, tornando necessárias pesquisas com intuito de melhorar a produtividade desse cereal. O objetivo desse trabalho foi a estimativa de produtividade do trigo, a partir da criação de modelos preditivos por meio da mineração de dados obtidos em imagens de alta resolução espacial, coletadas por aeronave remotamente pilotada (RPA). O trabalho foi realizado em duas áreas experimentais na cidade de Ponta Grossa – PR, onde para cada área foram feitas oito coletas de imagens, em diferentes estádios de desenvolvimento da cultura, com resolução espacial de 3,4cm/px e duas coletas com resolução 10cm/px e 20cm/px, através de uma RPA eBee utilizando uma câmera RGB e outra NIR. O processamento das imagens foi feito a partir do software Pix4D, e resultou em um ortomosaicos com os valores de refletância em diferentes comprimentos de onda: R, G e B da câmera RGB e R, G e NIR da câmera NIR, além de uma imagem com valores de NDVI obtidos a partir da aritmética das bandas Nir e Red (vermelho). A correção de georreferenciamento de cada ortomosaico e a extração dos valores de refletância foram feitas com auxílio do sistema de informação geográfica (SIG) Quantum GIS. A partir dos valores de refletância extraídos, foram criadas bases de dados em diferentes proporções (10%, 20%, 40%, 70% e 100%) para mineração de dados por meio do algoritmo SMOReg, baseado em máquina de vetor de suporte (SVM) para regressão (SVR). A correção de georreferenciamento utilizando 10 pontos de controle proporcionou ortomosaicos com erro médio quadrático (RSME) de distância de 0,35m, o qual não mostrou diferença significativa para a correção com 5 pontos de controle (RMSE = 0,38m). Os valores de refletância foram diferentes para cada área de estudo, tornando difícil a indicação de melhores períodos para a estimativa de produtividade do trigo. Os maiores índices de correlação da produtividade com os comprimentos de onda, foram obtidos com os dados das imagens da câmera RGB, seguido da câmera NIR e NDVI, com as correlações de 0,6168, 0,5423 e 0,5324, respectivamente. A quantidade de informação extraída das imagens, refletida na proporção das bases de dados, não se mostrou significativa nos modelos preditivos gerados, bem como nos índices de correlação, os quais foram estatisticamente iguais. Índices de correlação melhores foram obtidos a partir dos dados extraídos das imagens com resolução espacial de 20cm/px, o que sugere que imagens de alta resolução espacial podem não ser adequadas para estimativa de produtividade do trigo.