Orientador: Marco Antonio Garcia de Carvalho / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Tecnologia / Made available in DSpace on 2018-08-27T14:19:19Z (GMT). No. of bitstreams: 1 Rodrigues_LucasCerqueira_M.pdf: 3544177 bytes, checksum: ffa0c5e0ad4701affb1f2910bdd85ca4 (MD5) Previous issue date: 2015 / Resumo: Na pesquisa médica, o exame de sudorese é utilizado para destacar as regiões do corpo onde o paciente transpira, sendo estas úteis para o médico identificar possíveis lesões no sistema nervoso simpático. Os estudos acerca deste exame apontam a inexistência de um processo de identificação automática das regiões do corpo. Neste projeto, utilizou-se o Kinect® para ajudar nesta solução. Este dispositivo é capaz escanear objetos 3D e possui uma biblioteca para desenvolvimento de sistemas. Este trabalho tem o objetivo de construir um sistema computacional cujo propósito é desenvolver uma solução semi-automática para análise de imagens digitais provenientes de exames de sudorese. O sistema em foco permite classificar as regiões do corpo onde o paciente transpira, por intermédio de seu escaneamento 3D, utilizando o Kinect®, e gerar um relatório para o médico com as informações consolidadas de forma a realizar o diagnóstico com facilidade, rapidez e precisão. O projeto teve início em 2013, no laboratório IMAGELab da FT/UNICAMP em Limeira/SP e contou com o apoio de uma das equipes do Hospital das Clínicas da USP de Ribeirão Preto/SP que realiza os estudos sobre o Exame de Sudorese iodo-amido. A contribuição do trabalho consistiu na construção do aplicativo, que utiliza o algoritmo de segmentação de imagem K-Means para segmentação das regiões sobre a superfície do paciente, além do desenvolvimento do sistema que inclui o Kinect®. A aplicação validou-se por meio de experimentos em pacientes reais / Abstract: In medical research, the Sweat Test is used to highlight regions where the patient sweats, which are useful for the doctor to identify possible lesions on the sympathetic nervous system. Studies on this test indicate some difficulties in the automatic identification of body regions. In this project, we used the Kinect® device to help in this solution. Created by Microsoft®, the Kinect® is able to identify distance and has a library for systems development. This work aims to build a computer system intending to resolve some of the difficulties encountered during the research in the examination of sweating. The system created allows classify regions of the body where the patient sweats, through its 3D scanning, using the Kinect®, and export to the doctor the consolidated information in order to make a diagnosis quickly, easily and accurately. The project began in 2013 in ImageLab laboratory FT / UNICAMP in Limeira / SP and had the support of one of the USP Clinical Hospital teams in Ribeirão Preto / SP that performs studies on the Sweating Exam Iodine-Starch. The contribution to knowledge was in the software construction using the Kinect® and the image segmentation using K-Means algorithm for targeting regions on the surface of the patient. The application is validated by experiments on real patients / Mestrado / Tecnologia e Inovação / Mestre em Tecnologia

In this thesis, we introduce a radial basis function network approach to solve a color image classification problem in a real time industrial application. Radial basis function networks are employed to classify the images of finished wooden parts in terms of their color and species. Other classification methods are also examined in this work. The minimum distance classifiers are presented since they have been employed by the previous research. We give brief definitions about color space, color texture, color quantization, color classification methods. We also give an intensive review of radial basis functions, regularization theory, regularized radial basis function networks, and generalized radial basis function networks. The centers of the radial basis functions are calculated by the k-means clustering algorithm. We examine the k-means algorithm in terms of starting criteria, the movement rule, and the updating rule. The dilations of the radial basis functions are calculated using a statistical method. Learning classifier systems are also employed to solve the same classification problem. Learning classifier systems learn the training samples completely whereas they are not successful to classify the test samples. Finally, we present some simulation results for both radial basis function network method and learning classifier systems method. A comparison is given between the results of each method. The results show that the best classification method examined in this work is the radial basis function network method. / Master of Science

In this master thesis I recapitulated several methods for clustering input data. Two well known clustering algorithms, concretely K-means algorithm and Fuzzy C-means (FCM) algorithm, were described in the submitted work. I presented several methods, which could help estimate the optimal number of clusters. Further, I described Kohonen maps and two models of Kohonen's maps with dynamically changing structure, namely Kohonen map with growing grid and the model of growing neural gas. At last I described quite new model of radial basis function neural networks. I presented several learning algorithms for this model of neural networks. In the end of this work I made some clustering experiments with real data. This data describes the international trade among states of the whole world.

In this master thesis I recapitulated several methods for data clustering. Two well known clustering algorithms, concretely K-means algorithm and Fuzzy C-means (FCM) algorithm, were described in the submitted work. I presented several methods, which could help estimate the optimal number of clusters. Further, I described Kohonen maps and two models of Kohonen's maps with dynamically changing structure, namely Kohonen map with growing grid and the model of growing neural gas. At last I described quite new model of radial basis function neural networks. I presented several learning algorithms for this model of neural networks, RAN, RANKEF, MRAN, EMRAN and GAP. In the end of this work I made some clustering experiments with real data. This data describes the international trade among states of the whole world.

The programme for the segmentation of a speech into fonems was created as a part of the master´s thesis. This programme was made in the programme Matlab and consists of several scripts. The programme serves for automatic segmentation. Speech segmentation is the process of identifying the boundaries between phonemes in spoken natural languages. Automatic segmentation is based on vector quantization. In the first step of algorithm, feature extraction is realized. Then speech segments are assigned to calculated centroids. Position where centroid is changed is marked as a boundary of phoneme. The audiorecords were elaborated by the programme and a operation of the automatic segmentation was analysed. A detailed manual was created to the programme too. Individual used methods of the elaboration of a speech were in the master´s thesis briefly descripted, its implementations in the programme and reasons of set of its parameters.

This project is about trying to apply machine learning theories on a selection of data points in order to see if an improvement of current methodology within stress detection and measure selecting could be applicable for the company Linkura AB. Linkura AB is a medical technology company based in Linköping and handles among other things stress measuring for different companies employees, as well as health coaching for selecting measures. In this report we experiment with different methods and algorithms under the collective name of Unsupervised Learning, to identify visible patterns and behaviour of data points and further on we analyze it with the quantity of data received. The methods that have been practiced on during the project are “K-means algorithm” and a dynamic hierarchical clustering algorithm. The correlation between the different data points parameters is analyzed to optimize the resource consumption, also experiments with different number of parameters are tested and discussed with an expert in stress coaching. The results stated that both algorithms can create clusters for the risk groups, however, the dynamic clustering method clearly demonstrate the optimal number of clusters that should be used. Having consulted with mentors and health coaches regarding the analysis of the produced clusters, a conclusion that the dynamic hierarchical cluster algorithm gives more accurate clusters to represent risk groups were done. The conclusion of this project is that the machine learning algorithms that have been used, can categorize data points with stress behavioral correlations, which is usable in measure testimonials. Further research should be done with a greater set of data for a more optimal result, where this project can form the basis for the implementations. / Detta projekt handlar om att försöka applicera maskininlärningsmodeller på ett urval av datapunkter för att ta reda på huruvida en förbättring av nuvarande praxis inom stressdetektering och  åtgärdshantering kan vara applicerbart för företaget Linkura AB. Linkura AB är ett medicintekniskt företag baserat i Linköping och hanterar bland annat stressmätning hos andra företags anställda, samt hälso-coachning för att ta fram åtgärdspunkter för förbättring. I denna rapport experimenterar vi med olika metoder under samlingsnamnet oövervakad maskininlärning för att identifiera synbara mönster och beteenden inom datapunkter, och vidare analyseras detta i förhållande till den mängden data vi fått tillgodosett. De modeller som har använts under projektets gång har varit “K-Means algoritm” samt en dynamisk hierarkisk klustermodell. Korrelationen mellan olika datapunktsparametrar analyseras för att optimera resurshantering, samt experimentering med olika antal parametrar inkluderade i datan testas och diskuteras med expertis inom hälso-coachning. Resultaten påvisade att båda algoritmerna kan generera kluster för riskgrupper, men där den dynamiska modellen tydligt påvisar antalet kluster som ska användas för optimalt resultat. Efter konsultering med mentorer samt expertis inom hälso-coachning så drogs en slutsats om att den dynamiska modellen levererar tydligare riskkluster för att representera riskgrupper för stress. Slutsatsen för projektet blev att maskininlärningsmodeller kan kategorisera datapunkter med stressrelaterade korrelationer, vilket är användbart för åtgärdsbestämmelser. Framtida arbeten bör göras med ett större mängd data för mer optimerade resultat, där detta projekt kan ses som en grund för dessa implementeringar.

Digital growth and the adaption of internet-based solutions, particularly artificial intelligence and machine learning, have dramatically changed the way design is done today. This rapid change in technology has challenged the level of automation, which influences the human-automation interactions with the available colour-design tools (academic and commercial). As colour design and selection are known to be one of the most critical steps of any art or design journey, the currently available tools use one over the other approaches, from the automation-levels spectrum, when it comes to contextual search for colour palettes, colour-extracting, and colour compatibility. On the one hand, fully automated approaches could exclude the designers’ intervention; on the other hand, fully manual approaches could be affected by human errors and weaknesses. Both approaches tend to have problems when used in colour design tools, such as restricting the designers’ freedom, overwhelming designers with information-overload and option-widget clutter that exist in the interfaces of such tools, or limiting designers by the functionalities offered by the tool based on its purpose, causing it to partially support certain parts of the designers’ colour selection process rather than the whole process. The thesis focuses on investigating the possible solutions for balancing the automated and manual methods for generating colour palettes and supporting the designers’ non-standardised colour-selection processes while tailoring the solution to intellectually stimulate and engage designers who work in different design fields, in comparison with the Adobe Explore Page–which is one of the most well-known and established colour design tools in today’s market and one of the applications that offers a contextual search feature. To fulfil the purpose of this research, a web-based application was prototyped (named Paletto), which consists of the requirements for enabling the rapid generation and exploration of colour palette variations, supporting end-users to contextually search for palettes, and allowing users to apply constraints (via a preference selection list) for a holistic palette adjustment. Afterwards, the proposed application was evaluated with 20 individuals from the target audience, using both qualitative and quantitative approaches to prove the concept according to participants’ acceptance, estimate Paletto’s effectiveness on their workflow and design process, examine their engagement and experience when completing the exploratory tasks, and gather additional insights about the design or the conceptual design and implementation of the application. Paletto generally received positive responses towards (1) the accuracy and relevance of its search results, (2) the selection feature and its adaptability and flexibility for human interventions, and (3) the system’s feedback in terms of information accessibility (e.g., search word and number of pages in the pagination). However, the palette generation feature had partially negative responses where participants showed annoyance, confusion, and thought it was complicated. At the same time, several participants appreciated the diversity of the generated palettes and the conceptual design of Paletto in general. Paletto found to effectively facilitate the colour-selection process and designers’ workloads in several areas, such as: fulfilling the end-user goals of producing quality palettes to be used in design projects; resources-efficiency (e.g., money-preserving, effort facilitation, and time-saving) for inspirational image gathering; automatic colour extraction and palette generation; providing freedom and support of decision making to explore colour combinations and variations via the iterative preferences selection; supporting colour-pattern identification in the selections; providing variation and relevant results when searching inspirational image gathering with accurate colour extractions that represent the searched images. Moreover, Paletto proved to offer greater user engagement and a better user experience in comparison with Adobe’s Explore Page. This was due to the felt involvement and the continuous interactivity offered by Paletto’s search and preference-selection features that allowed iterative palette generation and modification. In conclusion, the evaluations indicated some pain-points and gaps in the current design that were discussed in this thesis, and are accordingly recommended to be investigated in future work.

This work is about suggestion of the software for neural network based image segmentation. It defines basic terms for this topics. It is focusing mainly at preperation imaging information for image segmentation using neural network. It describes and compares different aproaches for image segmentation.