Ramazan Tekin

1D-local binary pattern based feature extraction for classification of epileptic EEG signals

In this paper, an effective approach for the feature extraction of raw Electroencephalogram (EEG) signals by means of one-dimensional local binary pattern (1D-LBP) was presented. For the importance of making the right decision, the proposed method was performed to be able to get better features of the EEG signals. The proposed method was consisted of two stages: feature extraction by 1D-LBP and classification by classifier algorithms with features extracted. On the classification stage, the several machine learning methods were employed to uniform and non-uniform 1D-LBP features. The proposed method was also compared with other existing techniques in the literature to find out benchmark for an epileptic data set. The implementation results showed that the proposed technique could acquire high accuracy in classification of epileptic EEG signals. Also, the present paper is an attempt to develop a general-purpose feature extraction scheme, which can be utilized to extract features from different categories of EEG signals.

Evaluation of texture features for automatic detecting butterfly species using extreme learning machine

In this study, we present an application of extreme learning machine (ELM) and image processing techniques for identifying butterfly species as an alternative to conventional diagnostic methods. This paper evaluates the capability of butterfly species classification by using texture features of butterfly images. Two texture descriptors such as grey-level co-occurrence matrix (GLCM) and local binary patterns (LBP) were used for comparison purpose. ELM is employed for classification in butterfly-feature space. A total of 190 butterfly images belonging to 19 different species of Pieridae family were used. The identification accuracy of the proposed method was 98.25% and 96.45% with GLCM and LBP butterfly-feature spaces, respectively. The methodology presented herein effectively detected and classified these butterflies. These findings suggested that the proposed GLCM, LBP texture features extraction techniques and ELM algorithm are feasible and excellent in identification and classification of butterfly species.

Two novel local binary pattern descriptors for texture analysis

In this study, two novel local binary patterns were proposed.First one is based on spatial relations between neighbors with a distance parameter.The second is based on relations between a reference pixel and its neighbor on the same orientation.Two approaches are improved to detect special patterns in images.The results show that the proposed approaches can be used in image processing areas. The recent developments in the image quality, storage and data transmission capabilities increase the importance of texture analysis, which plays an important role in computer vision and image processing. Local binary pattern (LBP) is an effective statistical texture descriptor, which has successful applications in texture classification. In this paper, two novel descriptors were proposed to search different patterns in images built on LBP. One of them is based on the relations between the sequential neighbors with a specified distance and the other one is based on determining the neighbors in the same orientation through central pixel parameter. These descriptors are tested with the Brodatz-1, Brodatz-2, Butterfly and Kylberg datasets to show the applicability of the proposed nLBPd and dLBPα descriptors. The proposed methods are also compared with classical LBP. The average accuracies obtained by ANN with 10 fold cross validation, which are 99.26% (LBPu2 and nLBPd), 94.44% (dLBPα), 95.71% ( n L B P d u 2 ) and %99.64 (nLBPd), for Brodatz-1, Brodatz-2, Butterfly and Kylberg datasets, respectively, show that the proposed methods outperform significant accuracies.

Detection of Parkinson's disease by Shifted One Dimensional Local Binary Patterns from gait

This study showed that the PD can be diagnosed by using sensors attached at underfoot from gait.Feature extracted by Shifted 1D-LBP, which is sensitive to local changes in time signals.Shifted 1D-LBP has a simple algorithm. It can be used in real time applications.Obtained detection accuracy is 88.8889%.The accuracy results were compared with the results of previous studies in literature. The Parkinsons disease (PD) is one of the most common diseases, especially in elderly people. Although the previous studies showed that the PD can be diagnosed by expert systems through its cardinal symptoms such as the tremor, muscular rigidity, disorders of movements and voice, it was reported that the presented approaches, which utilize simple motor tasks, were limited and lack of standardization. To achieve a standard approach in PD detection, an approach, which is built on shifted one-dimensional local binary patterns (Shifted 1D-LBP) and machine learning methods, was proposed. Shifted 1D-LBP is built on 1D-LBP, which is sensitive to local changes in a signal. In 1D-LBP the positions of neighbors around center data are constant and therefore, the number of patterns that can be exacted by it is limited. This drawback was solved by Shifted 1D-LBP by changeable positions of neighbors. In evaluation and validation stages, the Gait in Parkinsons Disease (gaitpdb) dataset, which consists of three gait datasets that were recorded in different tasks or experiment protocols, were employed. Statistical features were exacted from formed histograms of gait signals transformed by Shifted 1D-LBP. Whole features and selected features were classified by machine learning methods. Obtained results were compared with statistical features exacted from signals in both time and frequency domains and results reported in the literature. Achieved results showed that the proposed approach can be successfully employed in PD detection from gait. This work is not only an attempt to develop a PD detection method, but also a general-purpose approach that is based on detecting local changes in time ordered signals.

A novel approach for SEMG signal classification with adaptive local binary patterns

AbstractFeature extraction plays a major role in the pattern recognition process, and this paper presents a novel feature extraction approach, adaptive local binary pattern (aLBP). aLBP is built on the local binary pattern (LBP), which is an image processing method, and one-dimensional local binary pattern (1D-LBP). In LBP, each pixel is compared with its neighbors. Similarly, in 1D-LBP, each data in the raw is judged against its neighbors. 1D-LBP extracts feature based on local changes in the signal. Therefore, it has high a potential to be employed in medical purposes. Since, each action or abnormality, which is recorded in SEMG signals, has its own pattern, and via the 1D-LBP these (hidden) patterns may be detected. But, the positions of the neighbors in 1D-LBP are constant depending on the position of the data in the raw. Also, both LBP and 1D-LBP are very sensitive to noise. Therefore, its capacity in detecting hidden patterns is limited. To overcome these drawbacks, aLBP was proposed. In aLBP, the positions of the neighbors and their values can be assigned adaptively via the down-sampling and the smoothing coefficients. Therefore, the potential to detect (hidden) patterns, which may express an illness or an action, is really increased. To validate the proposed feature extraction approach, two different datasets were employed. Achieved accuracies by the proposed approach were higher than obtained results by employed popular feature extraction approaches and the reported results in the literature. Obtained accuracy results were brought out that the proposed method can be employed to investigate SEMG signals. In summary, this work attempts to develop an adaptive feature extraction scheme that can be utilized for extracting features from local changes in different categories of time-varying signals.

The influence of ion concentrations on the dynamic behavior of the Hodgkin–Huxley model-based cortical network

Action potentials (APs) in the form of very short pulses arise when the cell is excited by any internal or external stimulus exceeding the critical threshold of the membrane. During AP generation, the membrane potential completes its natural cycle through typical phases that can be formatted by ion channels, gates and ion concentrations, as well as the synaptic excitation rate. On the basis of the Hodgkin–Huxley cell model, a cortical network consistent with the real anatomic structure is realized with randomly interrelated small population of neurons to simulate a cerebral cortex segment. Using this model, we investigated the effects of Na+ and K+ ion concentrations on the outcome of this network in terms of regularity, phase locking, and synchronization. The results suggested that Na+ concentration does slightly affect the amplitude but not considerably affects the other parameters specified by depolarization and repolarization. K+ concentration significantly influences the form, regularity, and synchrony of the network-generated APs. No previous study dealing directly with the effects of both Na+ and K+ ion concentrations on regularity and synchronization of the simulated cortical network-generated APs, allowing for the comparison of results obtained using our methods, was encountered in the literature. The results, however, were consistent with those obtained through studies concerning resonance and synchronization from another perspective and with the information revealed through physiological and pharmacological experiments concerning changing ion concentrations or blocking ion channels. Our results demonstrated that the regularity and reliability of brain functions have a strong relationship with cellular ion concentrations, and suggested the management of the dynamic behavior of the cellular network with ion concentrations.

EMG signal classification by extreme learning machine

From disease detection to action assessment EMG signals are used variety of field. Miscellaneous studies have been conducted toward analysis of EMG signals. In this study some statistical features of signal were derived, the best evocative features were selected via Linear Discriminant Analysis (LDA) and feature vectors were constructed. This analytic feature vectors were classified through Extreme Learning Machine (ELM). 8 channel EMG signals recorded from 10 normal and 10 aggressive actions were used as an example. By cross-comparison of the obtained results to the ones obtained via various feature identifying methods (AR coefficients, wavelet energy and entropy) and classification methods (NB, SVM, LR, ANN, PART, Jrip, J48 and LMT) the success of the proposed method was determined.

Effects of small-world rewiring probability and noisy synaptic conductivity on slow waves: Cortical network

Physiological rhythms play a critical role in the functional development of living beings. Many biological functions are executed with an interaction of rhythms produced by internal characteristics of scores of cells. While synchronized oscillations may be associated with normal brain functions, anomalies in these oscillations may cause or relate the emergence of some neurological or neuropsychological pathologies. This study was designed to investigate the effects of topological structure and synaptic conductivity noise on the spatial synchronization and temporal rhythmicity of the waves generated by cells in the network. Because of holding the ability of clustering and randomizing with change of parameters, small-world (SW) network topology was chosen. The oscillatory activity of network was tried out by manipulating an insulated SW, cortical network model whose morphology is very close to real world. According to the obtained results, it was observed that at the optimal probabilistic rates of conductivity noise and rewiring of SW, powerful synchronized oscillatory small waves are generated in relation to the internal dynamics of cells, which are in line with the network’s input. These two parameters were observed to be quite effective on the excitation-inhibition balance of the network. Accordingly, it may be suggested that the topological dynamics of SW and noisy synaptic conductivity may be associated with the normal and abnormal development of neurobiological structure.

Randomized feed-forward artificial neural networks in estimating short-term power load of a small house: A case study

Randomized feed-forward artificial neural networks (ANNs) have been employed in various domains. This paper was written in order to assess the efficiency of the basic forms of randomized feed-forward ANNs, which are randomized weight artificial neural network, random vector functional link network, extreme learning machine, and radial bases function neural network. In order to compare these methods, a complex dataset, which is the power load of a small house dataset, was used. Obtained results showed that lower training error rates were achieved by randomized vector functional link network. On the other hand, lower test error rates were achieved by ELM. Furthermore, ELM has faster training and test stages than the other employed randomized ANNs.

Classification of hand opening/closing and fingers by using two channel surface EMG signal

In this study, two-channel surface electromyogram (sEMG) signals were used to classify hand open/close with fingers. The bispectrum analysis of the sEMG signal recorded with surface electrodes near the region of the muscle bundles on the front and back of the forearm was classified by extreme learning machines (ELM) based on phase matches in the EMG signal. EMG signals belonging to 17 persons, 8 males and 9 females, with an average age of 24 were used in the study. The fingers were classified using ELM algorithm with 94.60% accuracy in average. From the information obtained through this study, it seems possible to control finger movements and hand opening/closing by using muscle activities of the forearm which we hope to lead to control of intelligent prosthesis hands with high degree of freedom.