Papers by Keyword: Classification

Abstract: Blindness usually comes from two main causes, glaucoma and diabetes. Robust mass screening is performed for diagnosing, such as screening that requires a cost-effective method for glaucoma and diabetic retinopathy and integrates well with digital medical imaging, image processing, and administrative processes. For addressing all these issues, we propose a novel low-cost automated glaucoma and diabetic retinopathy diagnosis system, based on features extraction from digital eye fundus images. This paper proposes a diagnosis system for automated identification of healthy, glaucoma, and diabetic retinopathy. Using a combination of local binary pattern features, Gabor filter features, statistical features, and color features which are then fed to an artificial neural network and support vector machine classifiers. In this work, the classifier identifies healthy, glaucoma, and diabetic retinopathy images with an accuracy of 91.1%,92.9%, 92.9%, and 92.3% and sensitivity of 91.06%, 92.6%, 92.66%, and 91.73% and specificity of 89.83%, 91.26%, 91.96%, and 89.16% for ANN, and an accuracy of 90.0%,92.94%, 95.43%, and 97.92% and sensitivity of 89.34%, 93.26%, 95.72%, and 97.93% and specificity of 95.13%, 96.68%, 97.88%, and 99.05% for SVM, based on 5, 10, 15, and 31 number of selected features. The proposed system can detect glaucoma, diabetic retinopathy and normal cases with high accuracy and sensitivity using selected features, the performance of the system is high due to using of a huge fundus database.

Abstract: The wavelet transform is an accurate, efficient and efficacious method to improve the quality of the myoelectric signal. Classification of the signal from upper limb using Surface Electromyogram (SEMG) signal has been the matter of extensive research. Number of methods and algorithms have been described by researchers to classify biomedical signals. The main aim of this paper to extract the different coefficient values from the given SEMG data by using Discrete Wavelet Transform (DWT). Afterward, random forest machine learning algorithm was used to identify the different shoulder movement of an upper limb amputee. The combination of wavelet coefficients and random forest exhibited the best performance with 99.2% accuracy for the classification of different shoulder motions. It was found that the different motion can be identified accurately and provide the fundamental information to develop an efficient prosthetic device.

Abstract: A high availability of machines has always been important in production. One way to increase it is to avoid unscheduled production stops by detecting the onset of machine faults and to conduct preventative repairs. The detection part consists of the three steps signal acquisition, feature extraction and classification. This paper focuses on the last two steps through the example of an induction motor. Based on a publicly available motor current data set, features were extracted using the continuous wavelet transform. In the subsequent classification step eight different classification methods were compared with each other. It was found, that the accuracy of the classifiers varied significantly in a range from 20.6 % to 92.8 %. Moreover, the supportive vector machine, scoring an accuracy of 92.8 %, was the only classifier with an accuracy above 55.0 %.

Abstract: The effect of fibre reinforcement on Load Induced Thermal Strains (LITS) has not yet been significantly investigated up to now. Creep is becoming a key research topic only in the last few years. A semi-empirical model able to take into account both the thermo-mechanical damage associated to coarse aggregates and the thermo-chemical damage induced in the matrix and calibrated on the basis of the main results on plain concrete available in the scientific literature is presented. Some tests in uniaxial compression on Fibre Reinforced Concrete (FRC) cylinders characterized by a long age – 11-years-old – have been investigated and compared with the model to highlight fibre effects, if any. The uniaxial compressive strength at 28 days of the SFRC was 75 MPa; the specimens after 11 years showed a compressive strength exceeding 110 MPa. A strong increase of SLS residual strength was observed in post-cracking tension due to the long aging, while ULS residual strengths weakly increased. The cylindrical specimens were exposed to a maximum temperature of 200°C and 400°C and loaded with two load thresholds corresponding to 20% and 40% of the compressive strength detected at 28 days of aging, that means about 12.5% and 25% of the 11-years-old specimens. Two paths were investigated: pre-heated specimens up to 200°C or 400°C, then loaded with a compression stress equal to 0.2f_c,28 and 0.4f_c,28; and pre-loaded specimens up to 0.2f_c,28 and 0.4f_c,28 and then heated up to 200°C or 400°C. The duration of each test did not exceed 12 hours. Two main fibre effects were observed: a significant reduction of irreversible strains when the specimens were loaded and then heated and cooled and a different evolution in LITS passing from 200°C to 400°C, characterized by a significant reduction of the expected deformation.

Abstract: Spam emails have become an increasing difficulty for the entire web-users.These unsolicited messages waste the resources of network unnecessarily. Customarily, machine learning techniques are adopted for filtering email spam. This article examines the capabilities of the extreme learning machine (ELM) and support vector machine (SVM) for the classification of spam emails with the class level (d). The ELM method is an efficient model based on single layer feed-forward neural network, which can choose weights from hidden layers,randomly. Support vector machine is a strong statistical learning theory used frequently for classification. The performance of ELM has been compared with SVM. The comparative study examines accuracy, precision, recall, false positive, true positive.Moreover, a sensitivity analysis has been performed by ELM and SVM for spam email classification.

Abstract: Significant attention is paid to the production of peat-based materials in modern days. The study explores the influence of natural peat thermal modification on its properties by applying class-modeling techniques. Modification of different types of peat is achieved by heating at 250 °C. The set of peat properties such as component composition, g-factor and IR-spectra is used to obtain data matrix. It is shown that class-modeling techniques, such as partial least-squares discriminant analysis (PLS-DA) and simple independent modeling of class analogy (SIMCA), allow estimating peat class (natural or modified) by a set of properties without prediction errors by using three latent variables. According to the results of classification, it is established that thermal modification can be considered as a means of regulating the composition and physico-chemical properties of natural peats as a raw material

Abstract: — Neural networks are frequently used as a classifier for tasks in many classifications. However there are disadvantages in terms of amount of training data required, and length of training time. This paper, develop an intelligent diagnosis system for zinc oxide (ZnO) surge arrester fault classification. First the features were extracted from 600 ZnO surge arrester thermal images and leakage currents. Then these features were presented to several neural network architectures to investigate the most suitable network model for classifying the ZnO surge arrester fault condition effectively. Three classification models were used namely feed forward back propagation (FFBP), radial basis function (RBF) and learning vector quantization (LVQ) algorithm. The performance of the networks was compared based on resulted of misclassify and correct rate. The method was evaluated using 24 testing datasets. Comparison results showed that LVQ was the best training algorithm for the ZnO surge arrester fault classification compared to the others system. Also the LVQ is faster than FFBP and RBF.

Abstract: — Thermal imaging technique is a very convenient, versatile and non-contact method which has been used for fault condition diagnosis of electrical equipment. The fault condition diagnosis is composed with data acquisition, data pre-processing, data analysis and decision making. Some important features contain in thermal image can be extracted for equipment condition monitoring and fault diagnosis. This paper attempts to extract some important features from the zinc oxide (ZnO) surge arrester using first order statistical histogram extraction to classify the fault condition using neural network. The experimental work was carried out to capture thermal image of 120 kV rated ZnO surge arrester. The thermal images were segmented and plotted histogram using dedicated software. Some features such as the maximum, minimum, mean, standard deviation, and variance were extracted using the extraction method, classification of aging was carried out using the neural network based on the leakage current values. The health states consist of normal, defection and faulty. The results show that the thermal image features extracted using the extraction method can be used to classify the fault condition of the ZnO surge arresters

Abstract: Bionics is a new frontier science and provides many inspiration and thoughts for human invention. Through discussing the classification of bionics, it can greatly improve the capacity of human’s adaptation for nature and productivity as well. It can also have great social and economic benefits. Besides, the reason that restricts bionics development is expounded. Development of bionics can’t last long without the cooperation with many other subjects, and its own development also promotes the development of these disciplines. Furthermore, it is development trend of bionic that to achieve more perfect imitation and reengineering of nature, to discover and develop its relevant theories and technical methods. Therefore, bionics plays more and more important role in innovation and has broad prospects in the research.

Abstract: The aim of this paper is to present the design of a condition monitoring tool, its use for the intelligent configuration of pattern recognition algorithms, for fault detection, and for diagnosis applications. The modular design and functionality of the tool will be introduced. The tool, developed and implemented by Fraunhofer IPK, can be used, in particular, to support the development process of algorithms for condition monitoring of wear-susceptible components in production systems. An example of the industrial application of the tool will be presented. This will include the implementation of configured algorithms using the tool on an embedded system using Raspberry Pi 2 and MEMS sensor. Finally, the evaluation of these algorithms on an axis test rig at different operating parameters will be presented.