As diversas áreas da Engenharia, em parceria com a ciência médica, têm contribuído de forma eficaz para o avanço do conhecimento e dos resultados em aplicações práticas na vida do deficiente físico. De forma geral, pesquisas com este foco têm permitido o desenvolvimento de dispositivos e recursos com o objetivo de oferecer novamente a mobilidade e a liberdade perdidas com a deficiência. Este trabalho tem a finalidade de desenvolver um sistema que utiliza Eletromiografia de Superfície e Máquina de Vetores de Suporte para a caracterização de determinados movimentos de um braço humano, possibilitando, futuramente, a integração em sistemas de reabilitação. Primeiramente os sinais mioelétricos são obtidos nos músculos do braço de voluntários através de eletrodos de superfície ligados a um eletromiógrafo. O sinal é adquirido, utilizando como padrão um modelo virtual que demonstra ao voluntário os movimentos do segmento mão-braço que devem ser imitados. Esses movimentos são executados e seus sinais mioelétricos adquiridos. Posteriormente, esses sinais são processados e características são extraídas. Em seguida, são alocadas algumas de suas características (RMS, média, variância, desvio padrão, skewness e kurtosis) na entrada da Máquina de Vetores de Suporte, que apresenta, como saída, o reconhecimento, ou não, do movimento previamente executado pelo voluntário. No final do processo, observou-se que aumentando o número de canais elevou-se a taxa de acerto dos movimentos e, com a retirada de determinada característica, houve decréscimo na taxa de acerto do sistema. Nestes casos, os 9 movimentos distintos atingiram uma taxa de acerto média de 83,2%, para dois canais, e 91,3%, para oito canais, e, em ambos sistemas de canais, com as seis características. / A wide range of engineering scopes, along with the knowledge from the medical science, has efficiently been contributing to further knowledge and results for practical applications in the life of the physically challenged. In general, such researches have allowed the development of devices and resources aimed at giving back the mobility and freedom lost with the deficiency. This paper intends to develop a system that uses Surface Electromyography and Support-Vector Machines (SVM) for the characterization of specific movements of a human arm enabling the future integration in rehabilitation systems. At first, myoelectric signals are obtained in the arm muscles of volunteers by means of surface electrodes attached to an Electromyography. The signal is acquired using a virtual model as pattern demonstrating to the volunteer the hand-arm movements which are to be replicated by the subject. As these movements are done, its respective myoelectric signals are acquired. Later on, these signals are processed and their characteristics extracted. Some of these features (such as RMS, standard deviation, variance, mean, kurtosis, skewness) will then be inserted in as input data in the Support- Vector Machine, which shows as an output a valid or null recognition of the movement earlier executed by the volunteer. At the end of the process, it was observed that increasing the number of channels increased by hit rate movements and, with the removal of certain characteristic, there was a decrease in the hit rate of the system. In these cases, nine distinct movements reached an average accuracy of 83.2% for two channels, and 91.3% for eight channels, and in both systems of channels, with six features.

As diversas áreas da Engenharia, em parceria com a ciência médica, têm contribuído de forma eficaz para o avanço do conhecimento e dos resultados em aplicações práticas na vida do deficiente físico. De forma geral, pesquisas com este foco têm permitido o desenvolvimento de dispositivos e recursos com o objetivo de oferecer novamente a mobilidade e a liberdade perdidas com a deficiência. Este trabalho tem a finalidade de desenvolver um sistema que utiliza Eletromiografia de Superfície e Máquina de Vetores de Suporte para a caracterização de determinados movimentos de um braço humano, possibilitando, futuramente, a integração em sistemas de reabilitação. Primeiramente os sinais mioelétricos são obtidos nos músculos do braço de voluntários através de eletrodos de superfície ligados a um eletromiógrafo. O sinal é adquirido, utilizando como padrão um modelo virtual que demonstra ao voluntário os movimentos do segmento mão-braço que devem ser imitados. Esses movimentos são executados e seus sinais mioelétricos adquiridos. Posteriormente, esses sinais são processados e características são extraídas. Em seguida, são alocadas algumas de suas características (RMS, média, variância, desvio padrão, skewness e kurtosis) na entrada da Máquina de Vetores de Suporte, que apresenta, como saída, o reconhecimento, ou não, do movimento previamente executado pelo voluntário. No final do processo, observou-se que aumentando o número de canais elevou-se a taxa de acerto dos movimentos e, com a retirada de determinada característica, houve decréscimo na taxa de acerto do sistema. Nestes casos, os 9 movimentos distintos atingiram uma taxa de acerto média de 83,2%, para dois canais, e 91,3%, para oito canais, e, em ambos sistemas de canais, com as seis características. / A wide range of engineering scopes, along with the knowledge from the medical science, has efficiently been contributing to further knowledge and results for practical applications in the life of the physically challenged. In general, such researches have allowed the development of devices and resources aimed at giving back the mobility and freedom lost with the deficiency. This paper intends to develop a system that uses Surface Electromyography and Support-Vector Machines (SVM) for the characterization of specific movements of a human arm enabling the future integration in rehabilitation systems. At first, myoelectric signals are obtained in the arm muscles of volunteers by means of surface electrodes attached to an Electromyography. The signal is acquired using a virtual model as pattern demonstrating to the volunteer the hand-arm movements which are to be replicated by the subject. As these movements are done, its respective myoelectric signals are acquired. Later on, these signals are processed and their characteristics extracted. Some of these features (such as RMS, standard deviation, variance, mean, kurtosis, skewness) will then be inserted in as input data in the Support- Vector Machine, which shows as an output a valid or null recognition of the movement earlier executed by the volunteer. At the end of the process, it was observed that increasing the number of channels increased by hit rate movements and, with the removal of certain characteristic, there was a decrease in the hit rate of the system. In these cases, nine distinct movements reached an average accuracy of 83.2% for two channels, and 91.3% for eight channels, and in both systems of channels, with six features.

Orientador: Ronei Jesus Poppi / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-08T20:32:52Z (GMT). No. of bitstreams: 1 Borin_Alessandra_D.pdf: 4356106 bytes, checksum: 6db61edddb59ead1ed109312c8c4cb58 (MD5) Previous issue date: 2007 / Resumo: Neste trabalho foram comparados modelos desenvolvidos utilizando as máquinas de vetores de suporte por mínimos quadrados (LS-SVM) e regressão por mínimos quadrados parciais (PLS) para três aplicações em determinação de parâmetros de qualidade em matrizes lácteas. Na primeira aplicação determinaram-se simultaneamente três adulterantes em leite em pó (amido, soro de leite e sacarose), empregando espectroscopia no infravermelho próximo por reflectância difusa. Em todos os modelos desenvolvidos, os resultados obtidos com LS-SVM foram superiores ao PLS, apresentando erros de previsão na faixa de 1,10% para amido, 1,10% para soro e 0,60% para sacarose, comparados aos erros de previsão obtidos pelo PLS: 2,40% para amido, 4,00% para soro e 3,00% para sacarose. Na segunda aplicação, quantificou-se Lactobacillus em leite fermentado a partir de imagens escaneadas de placas de crescimento microbiológico. A imagem foi convertida em histograma de cores, para ser usada no desenvolvimento dos modelos PLS e LS-SVM. O modelo desenvolvido pelo LS-SVM apresentou erros de previsão da ordem de 22,16 ufc e o PLS apresentou erros de previsão de 41,46 ufc. Na última aplicação, foi realizada a determinação do índice de gordura em leite humano, provenientes de diferentes doadoras, empregando a espectroscopia no infravermelho médio por reflectância total atenuada. Novamente o modelo LS-SVM apresentou melhores resultados em relação ao PLS, onde para a LS-SVM foram obtidos erros de previsão da ordem de 41,00 Kcal/L e, para o PLS, erros de previsão de 103,00 Kcal/L. / Abstract: In this work calibration models by using Least-square support vector machine (LS-SVM) and Partial least square (PLS) regression were compared for three applications in determination of quality parameters in milky matrices. In the first application, three adulterants in powdered milk (starch, whey and sucrose) had been simultaneously determined employing near infrared spectroscopy by diffuse reflectance. In all developed models the results obtained with LS-SVM were superior to PLS, presenting prediction errors of 1,10% for starch, 1,10% for whey and 0,60% for sucrose compared with the errors obtained by the PLS of 2,40% for starch, 4,00% for whey and 3,00% for sucrose. In the second application, Lactobacillus in fermented milk were quantified from digital images of microbiological plates of growth. The images was converted into histogram of colors to be used in the development of PLS and LS-SVM models. The model developed for the LS-SVM presented errors of 22,16 cfu an the PLS presented prediction errors of 41,46 cfu. In the last application, the fat index in human milk was determined, proceeding from different mothers, employing Mid spectroscopy by attenuated total reflectance. Again, LS-SVM model presented better reulted than PLS, with prediction error 41,00 Kcal/L against PLS the prediction error of 103,00 Kcal/L. / Doutorado / Quimica Analitica / Doutor em Ciências

Made available in DSpace on 2016-08-17T14:52:37Z (GMT). No. of bitstreams: 1 Dissertacao Suellen de Araujo Caduda da Silva Motta.pdf: 13615714 bytes, checksum: 2d717995db073f7cf237e113b96957a4 (MD5) Previous issue date: 2015-02-02 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / This work presents an automatic method for fault detection in data obtained through seismic reflection method. Identifying geological faults in seismic data is critical for better understating a geological system and planning hydrocarbon exploration. Knowing that faults are discontinuities present in seismic horizons, we propose the use of geostatistical functions which are capable of indicating the amplitude variation along the volume samples, in both predetermined distances and directions. Thus, the method is based on semivariogram, semimadogram, covariogram and correlogram functions, used as representative characteristics for the samples, which will be classified as fault or "non fault" regions by the Pattern Recognition technique named Support Vector Machine (SVM). The proposed method was validated by tests made in F3 Block, a seismic data provided by OpendTect system, with up to 92.15% sensitivity and 84.33% specificity. This work also provides an extraction of fault lines method based on region growing segmentation and morphological operators applied on the classification binary resulted volume. Also tested in F3 Block, the method was able to satisfactorily extract the faults in most of the data slices. / Este trabalho apresenta um método automático de detecção de falhas em volumes obtidos através do método de reflexão sísmica. Identificar as falhas geológicas nos dados sísmicos é importante para o conhecimento de um sistema geológico e para o planejamento da exploração de hidrocarbonetos. Sabendo-se que as falhas são descontinuidades presentes nos horizontes sísmicos, propõe-se a utilização de funções geoestatísticas capazes de indicar a variação da amplitude das amostras, em direções e distâncias predeterminadas. Assim, o método baseia-se no uso das funções semivariograma, semimadograma, covariograma e correlograma como características representativas das amostras, que serão classificadas como regiões de falha ou não falha , através da técnica clássica de Reconhecimento de Padrões conhecida como SVM (Support Vector Machine Máquina de Vetores de Suporte). O método proposto foi validado através de testes realizados com o volume F3 Block, disponibilizado pelo sistema OpendTect, apresentando até 92,15% de sensibilidade e 84,33% de especificidade. Este trabalho também apresenta um método de extração das linhas de falha baseado em crescimento de região e operadores morfológicos, a partir do volume binário resultante da classificação. Também testado sobre o F3 Block, o método foi capaz de extrair satisfatoriamente as falhas, na maioria das fatias do dado.

Text categorisation is an important feature for organising text data and making it easier to find information on the world wide web.  The categorisation of text data can be done through the use of machine learning classifiers. These classifiers need to be trained with data in order to predict a result for future input. The authors chose to investigate how accurate two classifiers are when classifying recent news articles on a classifier model that is trained with older news articles. To reach a result the authors chose the Naive Bayes and Support Vector Machine classifiers and conducted an experiment. The experiment involved training models of both classifiers with news articles from 2005 and testing the models with news articles from 2005 and 2017 to compare the results. The results showed that both classifiers did considerably worse when classifying the news articles from 2017 compared to classifying the news articles from the same year as the training data.

With the growing phenomenon of an aging population, an increasing numberof older people are living alone for domestic and social reasons. Based on thisfact, falling accidents become one of the most important factors in threateningthe lives of the elderly. Therefore, it is necessary to set up an application to de-tect the daily activities of the elderly. However, falling detection is difficult to recognize because the "falling" motion is an instantaneous motion and easy to confuse with others.In this thesis, three data mining methods were employed on wearable sensors' value; first which contains the continuous data set concerning eleven activities of daily living, and then an analysis of the different results was performed. Not only could the fall be detected, but other activities could also be classified. In detail, three methods including Back Propagation Neural Network, Support Vector Machine and Hidden Markov Model are applied separately to train the data set.What highlights the project is that a new  idea is put forward, the aim of which is to design a methodology of accurate classification in the time-series data set. The proposed approach, which includes obtaining of classifier parts and the application parts allows the generalization of classification. The preliminary results indicate that the new method achieves the high accuracy of classification,and significantly performs better than other data mining methods in this experiment.

>Magister Scientiae - MSc / The South African Sign Language research group at the University of the Western Cape is in the process of creating a fully-edged machine translation system to automatically translate between South African Sign Language and English. A major component of the system is the ability to accurately recognise facial expressions, which are used to convey emphasis, tone and mood within South African Sign Language sentences. Traditionally, facial expression recognition research has taken one of two paths: either recognising whole facial expressions of which there are six i.e. anger, disgust, fear, happiness, sadness, surprise, as well as the neutral expression; or recognising the fundamental components of facial expressions as defined by the Facial Action Coding System in the form of Action Units. Action Units are directly related to the motion of specific muscles in the face, combinations of which are used to form any facial expression. This research investigates enhanced recognition of whole facial expressions by means of a hybrid approach that combines traditional whole facial expression recognition with Action Unit recognition to achieve an enhanced classification approach.

Number of lanes is a basic roadway attribute that is widely used in many transportation applications. Traditionally, number of lanes is collected and updated through field surveys, which is expensive especially for large coverage areas with a high volume of road segments. One alternative is through manual data extraction from high-resolution aerial images. However, this is feasible only for smaller areas. For large areas that may involve tens of thousands of aerial images and millions of road segments, an automatic extraction is a more feasible approach. This dissertation aims to improve the existing process of extracting number of lanes from aerial images automatically by making improvements in three specific areas: (1) performance of lane model, (2) automatic acquisition of external knowledge, and (3) automatic lane location identification and reliability estimation. In this dissertation, a framework was developed to automatically recognize and extract number of lanes from geo-rectified aerial images. In order to address the external knowledge acquisition problem in this framework, a mapping technique was developed to automatically estimate the approximate pixel locations of road segments and the travel direction of the target roads in aerial images. A lane model was developed based on the typical appearance features of travel lanes in color aerial images. It provides more resistance to “noise” such as presence of vehicle occlusions and sidewalks. Multi-class classification test results based on the K-nearest neighbor, logistic regression, and Support Vector Machine (SVM) classification algorithms showed that the new model provides a high level of prediction accuracy. Two optimization algorithms based on fixed and flexible lane widths, respectively, were then developed to extract number of lanes from the lane model output. The flexible lane-width approach was recommended because it solved the problems of error-tolerant pixel mapping and reliability estimation. The approach was tested using a lane model with two SVM classifiers, i.e., the Polynomial kernel and the Radial Basis Function (RBF) kernel. The results showed that the framework yielded good performance in a general test scenario with mixed types of road segments and another test scenario with heavy plant occlusions.

In the industry, a lot of companies are facing the explosion of big data. With this much information stored, companies want to make sense of the data and use it to help them for better decision making, especially for future prediction. A lot of money can be saved and huge revenue can be generated with the power of big data. When building statistical learning models for prediction, companies in the industry are aiming to build models with efficiency and high accuracy. After the learning models have been developed for production, new data will be generated. With the updated data, the models have to be updated as well. Due to this nature, the model performs best today doesn’t mean it will necessarily perform the same tomorrow. Thus, it is very hard to decide which algorithm should be used to build the learning model. This paper introduces a new method that ensembles the information generated by two different classification statistical learning algorithms together as inputs for another learning model to increase the final prediction power. The dataset used in this paper is NASA’s Turbofan Engine Degradation data. There are 49 numeric features (X) and the response Y is binary with 0 indicating the engine is working properly and 1 indicating engine failure. The model’s purpose is to predict whether the engine is going to pass or fail. The dataset is divided in training set and testing set. First, training set is used twice to build support vector machine (SVM) and neural network models. Second, it used the trained SVM and neural network model taking X of the training set as input to predict Y1 and Y2. Then, it takes Y1 and Y2 as inputs to build the Penalized Logistic Regression model, which is the ensemble model here. Finally, use the testing set follow the same steps to get the final prediction result. The model accuracy is calculated using overall classification accuracy. The result shows that the ensemble model has 92% accuracy. The prediction accuracies of SVM, neural network and ensemble models are compared to prove that the ensemble model successfully captured the power of the two individual learning model.

Recently, automatic demographic classification has found its way into embedded applications such as targeted advertising in mobile devices, and in-car warning systems for elderly drivers. In this thesis, we present a complete framework for video-based gender classification and age estimation which can perform accurately on embedded systems in real-time and under unconstrained conditions. We propose a segmental dimensionality reduction technique utilizing Enhanced Discriminant Analysis (EDA) to minimize the memory and computational requirements, and enable the implementation of these classifiers for resource-limited embedded systems which otherwise is not achievable using existing resource-intensive approaches. On a multi-resolution feature vector we have achieved up to 99.5% compression ratio for training data storage, and a maximum performance of 20 frames per second on an embedded Android platform. Also, we introduce several novel improvements such as face alignment using the nose, and an illumination normalization method for unconstrained environments using bilateral filtering. These improvements could help to suppress the textural noise, normalize the skin color, and rectify the face localization errors. A non-linear Support Vector Machine (SVM) classifier along with a discriminative demography-based classification strategy is exploited to improve both accuracy and performance of classification. We have performed several cross-database evaluations on different controlled and uncontrolled databases to assess the generalization capability of the classifiers. Our experiments demonstrated competitive accuracies compared to the resource-demanding state-of-the-art approaches.