Automated visual inspection is an emerging multi-disciplinary field with many challenges; it combines different aspects of computer vision, pattern recognition, automation, and control systems. There does not exist a large body of work dedicated to the design of generalized visual inspection systems; that is, those that might easily be made applicable to different product types. This is an important oversight, in that many improvements in design and implementation times, as well as costs, might be realized with a system that could easily be made to function in different production environments. <br /><br /> This thesis proposes a framework for generalizing and automating the design of the defect classification stage of an automated visual inspection system. It involves using an expandable set of features which are optimized along with the classifier operating on them in order to adapt to the application at hand. The particular implementation explored involves optimizing the feature set in disjoint sets logically grouped by feature type to keep search spaces reasonable. Operator input is kept at a minimum throughout this customization process, since it is limited only to those cases in which the existing feature library cannot adequately delineate the classes at hand, at which time new features (or pools) may have to be introduced by an engineer with experience in the domain. <br /><br /> Two novel methods are put forward which fit well within this framework: cluster-space and hybrid-space classifiers. They are compared in a series of tests against both standard benchmark classifiers, as well as mean and majority vote multi-classifiers, on feature sets comprised of just the logical feature subsets, as well as the entire feature sets formed by their union. The proposed classifiers as well as the benchmarks are optimized with both a progressive combinatorial approach and with an genetic algorithm. Experimentation was performed on true colour industrial lumber defect images, as well as binary hand-written digits. <br /><br /> Based on the experiments conducted in this work, it was found that the sequentially optimized multi hybrid-space methods are capable of matching the performances of the benchmark classifiers on the lumber data, with the exception of the mean-rule multi-classifiers, which dominated most experiments by approximately 3% in classification accuracy. The genetic algorithm optimized hybrid-space multi-classifier achieved best performance however; an accuracy of 79. 2%. <br /><br /> The numeral dataset results were less promising; the proposed methods could not equal benchmark performance. This is probably because the numeral feature-sets were much more conducive to good class separation, with standard benchmark accuracies approaching 95% not uncommon. This indicates that the cluster-space transform inherent to the proposed methods appear to be most useful in highly dependant or confusing feature-spaces, a hypothesis supported by the outstanding performance of the single hybrid-space classifier in the difficult texture feature subspace: 42. 6% accuracy, a 6% increase over the best benchmark performance. <br /><br /> The generalized framework proposed appears promising, because classifier performance over feature sets formed by the union of independently optimized feature subsets regularly met and exceeded those classifiers operating on feature sets formed by the optimization of the feature set in its entirety. This finding corroborates earlier work with similar results [3, 9], and is an aspect of pattern recognition that should be examined further.

Automated visual inspection is an emerging multi-disciplinary field with many challenges; it combines different aspects of computer vision, pattern recognition, automation, and control systems. There does not exist a large body of work dedicated to the design of generalized visual inspection systems; that is, those that might easily be made applicable to different product types. This is an important oversight, in that many improvements in design and implementation times, as well as costs, might be realized with a system that could easily be made to function in different production environments. <br /><br /> This thesis proposes a framework for generalizing and automating the design of the defect classification stage of an automated visual inspection system. It involves using an expandable set of features which are optimized along with the classifier operating on them in order to adapt to the application at hand. The particular implementation explored involves optimizing the feature set in disjoint sets logically grouped by feature type to keep search spaces reasonable. Operator input is kept at a minimum throughout this customization process, since it is limited only to those cases in which the existing feature library cannot adequately delineate the classes at hand, at which time new features (or pools) may have to be introduced by an engineer with experience in the domain. <br /><br /> Two novel methods are put forward which fit well within this framework: cluster-space and hybrid-space classifiers. They are compared in a series of tests against both standard benchmark classifiers, as well as mean and majority vote multi-classifiers, on feature sets comprised of just the logical feature subsets, as well as the entire feature sets formed by their union. The proposed classifiers as well as the benchmarks are optimized with both a progressive combinatorial approach and with an genetic algorithm. Experimentation was performed on true colour industrial lumber defect images, as well as binary hand-written digits. <br /><br /> Based on the experiments conducted in this work, it was found that the sequentially optimized multi hybrid-space methods are capable of matching the performances of the benchmark classifiers on the lumber data, with the exception of the mean-rule multi-classifiers, which dominated most experiments by approximately 3% in classification accuracy. The genetic algorithm optimized hybrid-space multi-classifier achieved best performance however; an accuracy of 79. 2%. <br /><br /> The numeral dataset results were less promising; the proposed methods could not equal benchmark performance. This is probably because the numeral feature-sets were much more conducive to good class separation, with standard benchmark accuracies approaching 95% not uncommon. This indicates that the cluster-space transform inherent to the proposed methods appear to be most useful in highly dependant or confusing feature-spaces, a hypothesis supported by the outstanding performance of the single hybrid-space classifier in the difficult texture feature subspace: 42. 6% accuracy, a 6% increase over the best benchmark performance. <br /><br /> The generalized framework proposed appears promising, because classifier performance over feature sets formed by the union of independently optimized feature subsets regularly met and exceeded those classifiers operating on feature sets formed by the optimization of the feature set in its entirety. This finding corroborates earlier work with similar results [3, 9], and is an aspect of pattern recognition that should be examined further.

Increasing power demand and wide use of high technology power electronic devices result in need for power quality monitoring. The quality of electric power in both transmission and distribution systems should be analyzed in order to sustain power system reliability and continuity. This analysis is possible by examination of data collected by power quality monitoring systems. In order to define the characteristics of the power system and reveal the relations between the power quality events, huge amount of data should be processed. In this thesis, clustering methods for power quality events are developed using exclusive and overlapping clustering models. The methods are designed to cluster huge amount of power quality data which is obtained from the online monitoring of the Turkish Electricity Transmission System. The main issues considered in the design of the clustering methods are the amount of the data, efficiency of the designed algorithm and queries that should be supplied to the domain experts. This research work is fully supported by the Public Research grant Committee (KAMAG) of TUBITAK within the scope of National Power quality Project (105G129).

Transitions between normal and pathological brain states are manifested differently in the electroencephalogram (EEG). Traditional discrimination of these states is often subject to bias and strict definitions. A fuzzy logic-based analysis can permit the classification and tracking of brain states in a non-subjective and unsupervised manner. In this thesis, the combination of fuzzy c-means (FCM) clustering, wavelet, and information theory has revealed notable frequency features in epilepsy and anaesthetic-induced unconsciousness. It was shown that entropy changes in membership functions correlate to specific epileptiform activity and changes in anaesthetic dosages. Seizure episodes appeared in the 31-39 Hz band, suggesting changes in cortical functional organization. The induction of anaesthetics appeared in the 64-72 Hz band, while the return to consciousness appeared in the 32-40 Hz band. Changes in FCM activity were associated with the concentration of anaesthetics. These results can help with the treatment of epilepsy and the safe administration of anaesthesia.

Transitions between normal and pathological brain states are manifested differently in the electroencephalogram (EEG). Traditional discrimination of these states is often subject to bias and strict definitions. A fuzzy logic-based analysis can permit the classification and tracking of brain states in a non-subjective and unsupervised manner. In this thesis, the combination of fuzzy c-means (FCM) clustering, wavelet, and information theory has revealed notable frequency features in epilepsy and anaesthetic-induced unconsciousness. It was shown that entropy changes in membership functions correlate to specific epileptiform activity and changes in anaesthetic dosages. Seizure episodes appeared in the 31-39 Hz band, suggesting changes in cortical functional organization. The induction of anaesthetics appeared in the 64-72 Hz band, while the return to consciousness appeared in the 32-40 Hz band. Changes in FCM activity were associated with the concentration of anaesthetics. These results can help with the treatment of epilepsy and the safe administration of anaesthesia.

With growing demands for reliability, availability, safety and cost efficiency in modern machinery, accurate fault diagnosis is becoming of paramount importance so that potential failures can be better managed. Although various methods have been applied to machinery condition monitoring and fault diagnosis, the diagnostic accuracy that can be attained is far from satisfactory. As most machinery faults lead to increases in vibration levels, vibration monitoring has become one of the most basic and widely used methods to detect machinery faults. However, current vibration monitoring methods largely depend on signal processing techniques. This study is based on the recognition that a multi-parameter data fusion approach to diagnostics can produce more accurate results. Fuzzy measures and fuzzy integral data fusion theory can represent the importance of each criterion and express certain interactions among them. This research developed a novel, systematic and effective fuzzy measure and fuzzy integral data fusion approach for machinery fault diagnosis, which comprises feature set selection schema, feature level data fusion schema and decision level data fusion schema for machinery fault diagnosis. Different feature selection and fault diagnostic models were derived from these schemas. Two fuzzy measures and two fuzzy integrals were employed: the 2-additive fuzzy measure, the fuzzy measure, the Choquet fuzzy integral and the Sugeno fuzzy integral respectively. The models were validated using rolling element bearing and electrical motor experiments. Different features extracted from vibration signals were used to validate the rolling element bearing feature set selection and fault diagnostic models, while features obtained from both vibration and current signals were employed to assess electrical motor fault diagnostic models. The results show that the proposed schemas and models perform very well in selecting feature set and can improve accuracy in diagnosing both the rolling element bearing and electrical motor faults.

To be successful in today&apos;s competitive market, service providers should look at user&apos;s satisfaction as a critical key. In order to gain a better understanding of customers&apos; expectations, a proper evaluations which considers intrinsic characteristics of perceived quality of service is needed. Due to the subjective nature of quality, the vagueness of human judgment and the uncertainty about the degree of users&apos; linguistic satisfaction, fuzziness is associated with quality of experience. Considering the capability of Fuzzy logic in dealing with imprecision and qualitative knowledge, it would be wise to apply it as a powerful mathematical tool for analyzing the quality of experience (QoE). This thesis proposes a fuzzy procedure to evaluate the quality of experience. In our proposed methodology, we provide a fuzzy relationship between QoE and Quality of Service (QoS) parameters. To identify this fuzzy relationship a new term called Fuzzi ed Opinion Score (FOS) representing a fuzzy quality scale is introduced. A fuzzy data mining method is applied to construct the required number of fuzzy sets. Then, the appropriate membership functions describing fuzzy sets are modeled and compared with each other. The proposed methodology will assist service providers for better decision-making and resource management.

This thesis is focused on cluster analysis in the field of text mining and its application to real data. The aim of the thesis is to find suitable categories (clusters) in the transcribed calls recorded in the contact center of Česká pojišťovna a.s. by transferring these textual documents into the vector space using basic text mining methods and the implemented clustering algorithms. From the formal point of view, the thesis contains a description of preprocessing and representation of textual data, a description of several common clustering methods, cluster validation, and the application itself.

Traumatic brain injury is highly prevalent in the United States yet there is little understanding of how the brain responds during injurious loading. A confounding problem is that because testing conditions vary between assessment methods, brain biomechanics cannot be fully understood. Data mining techniques were applied to discover how changes in testing conditions affect the mechanical response of the brain. Data were gathered from literature sources and self-organizing maps were used to conduct a sensitivity analysis to rank considered parameters by importance. Fuzzy C-means clustering was applied to find any data patterns. The rankings and clustering for each data set varied, indicating that the strain rate and type of deformation influence the role of these parameters. Multivariate linear regression was applied to develop a model which can predict the mechanical response from different experimental conditions. Prediction of response depended primarily on strain rate, frequency, brain matter composition, and anatomical region.

Yes / In dealing with big data, we need effective algorithms; effectiveness that depends, among others, on the ability to remove outliers from the data set, especially when dealing with classification problems. To this aim, support vector finder algorithms have been created to save just the most important data in the data pool. Nevertheless, existing classification algorithms, such as Fuzzy C-Means (FCM), suffer from the drawback of setting the initial cluster centers imprecisely. In this paper, we avoid existing shortcomings and aim to find and remove unnecessary data in order to speed up the final classification task without losing vital samples and without harming final accuracy; in this sense, we present a unique approach for finding support vectors, named evolutionary Pentagon Support Vector (PSV) finder method. The originality of the current research lies in using geometrical computations and evolutionary algorithms to make a more effective system, which has the advantage of higher accuracy on some data sets. The proposed method is subsequently tested with seven benchmark data sets and the results are compared to those obtained from performing classification on the original data (classification before and after PSV) under the same conditions. The testing returned promising results.