Classification of EMG signals using PSO optimized SVM for diagnosis of neuromuscular disorders

Support vector machines (SVMs) have been widely used in many pattern recognition problems. Generally, the performance of SVM classifiers is affected by the selection of the kernel parameters. However, SVM does not offer the mechanism for proper setting of their control parameters. The objective of this research is to optimize the parameters without degrading the SVM classification accuracy in diagnosis of neuromuscular disorders. An evolutionary approach for designing an SVM-based classifier (ESVM) by optimization of automatic parameter tuning using genetic algorithm is proposed. To illustrate and evaluate the efficiency of ESVM, a typical application to EMG signals classification using normal, myopathic, and neurogenic datasets is adopted. In the proposed method, the EMG signals were decomposed into the frequency sub-bands using discrete wavelet transform (DWT), and a set of statistical features was extracted from the sub-bands to represent the distribution of wavelet coefficients. It is shown that ESVM can obtain a high accuracy of 97 % using tenfold cross-validation for the EMG datasets. ESVM is developed as an efficient tool, so that various SVMs can be used conveniently as the core of ESVM for diagnosis of neuromuscular disorders.

In this paper, the features based on Intrinsic Mode Functions (IMFs) for classification of EMG signals are presented. The EMD method decomposes EMG signals into a set of narrow-band components known as IMFs. The features, namely mean frequency estimation and singular value computation, extracted from IMFs are exploited for classification of EMG signals. These parameters are used as an input to Least Squares Support Vector Machine (LS-SVM) with Radial Basis Function (RBF) for automatic classification of EMG signals. The classification accuracy for classification of normal and abnormal EMG signals obtained by the proposed method is 99.03% with RBF kernel of LS-SVM. The experimental results are presented to show the effectiveness of the proposed method for classification of normal and abnormal EMG signals (myopathy and neuropathy).

In this study, EMG signals taken from the skin surface as a result of muscles' contraction are classified. Studied EMG signals include 400 different patterns relating to four different movements. Each pattern is obtained by adding EMG signals one after another, which are recorded synchronously from two different muscles relating to one movement. Support Vector Machine (SVM) classifier, a supervised method, is used to classify these pattterns. But signals need to be preprocessed before being used in SVM classifier. To this end, spectral methods are consulted. In this way, feature vectors which are more significant than raw data and are composed of coefficients are achieved. Four different methods are used for preprocessing and feature vectors obtained are classified by SVM. Success of SVM classifier is tested and performances of preprocessing methods are compared. Best achievement is 94.25%. Keywords: EMG; Spectral Methods; Autoregressive (AR); SVM Classifier.

BackgroundSurface electromyography (sEMG) signals have been used in numerous studies for the classification of hand gestures and movements and successfully implemented in the position control of different prosthetic hands for amputees. sEMG could also potentially be used for controlling wearable devices which could assist persons with reduced muscle mass, such as those suffering from sarcopenia. While using sEMG for position control, estimation of the intended torque of the user could also provide sufficient information for an effective force control of the hand prosthesis or assistive device. This paper presents the use of pattern recognition to estimate the torque applied by a human wrist and its real-time implementation to control a novel two degree of freedom wrist exoskeleton prototype (WEP), which was specifically developed for this work.MethodsBoth sEMG data from four muscles of the forearm and wrist torque were collected from eight volunteers by using a custom-made testing rig. The features that were extracted from the sEMG signals included root mean square (rms) EMG amplitude, autoregressive (AR) model coefficients and waveform length. Support Vector Machines (SVM) was employed to extract classes of different force intensity from the sEMG signals. After assessing the off-line performance of the used classification technique, the WEP was used to validate in real-time the proposed classification scheme.ResultsThe data gathered from the volunteers were divided into two sets, one with nineteen classes and the second with thirteen classes. Each set of data was further divided into training and testing data. It was observed that the average testing accuracy in the case of nineteen classes was about 88% whereas the average accuracy in the case of thirteen classes reached about 96%. Classification and control algorithm implemented in the WEP was executed in less than 125 ms.ConclusionsThe results of this study showed that classification of EMG signals by separating different levels of torque is possible for wrist motion and the use of only four EMG channels is suitable. The study also showed that SVM classification technique is suitable for real-time classification of sEMG signals and can be effectively implemented for controlling an exoskeleton device for assisting the wrist.

Recently Autosomal Recessive Single Gene (ARSG) diseases are highly effective to the children within the age of 5–10 years. One of the most ARSG disease is a Phenylketonuria (PKU). This single gene disease is associated with mutations in the gene that encodes the enzyme phenylalanine hydroxylase (PAH, Gene 612349). Through this mutation process, PAH of the gene affected patient can not properly manufacture PAH as a result the patients suffer from decreased muscle tone which shows abnormality in EMG signal. Here the extraction of the quality of the PKU affected EMG (PKU-EMG) signal is a keen interest, so it is highly necessary to remove the added ECG signal as well as the biological and instrumental noises. In the Present paper we proposed a method for detection and classification of the PKU affected EMG signal. Here Discrete Wavelet Transformation is implemented for extraction of the features of the PKU affected EMG signal. Adaptive Neuro-Fuzzy Inference System (ANFIS) network is used for the classification of the signal. Modified Particle Swarm Optimization (MPSO) and Modified Genetic Algorithm (MGA) are used to train the ANFIS network. Simulation result shows that the proposed method gives better performance as compared to existing approaches. Also it gives better accuracy of 98.02% for the detection of PKU-EMG signal. The advantages of the proposed model is to use MGA and MPSO to train the parameters of ANFIS network for classification of ECG and EMG signal of PKU affected patients. The proposed method obtained the high SNR (18.13 ± 0.36 dB), SNR (0.52 ± 1.62 dB), RE (0.02 ± 0.32), MSE (0.64 ± 2.01), CC (0.99 ± 0.02), RMSE (0.75 ± 0.35) and MFRE (0.01 ± 0.02), RMSE (0.75 ± 0.35) and MFRE (0.01 ± 0.02). From authors knowledge, this is the first time a composite method is used for diagnosis of PKU affected patients. The accuracy (98.02%), sensitivity (100%) and specificity (98.59%) helps for proper clinical treatment. It can help for readers/researchers to improve the aforesaid performance for future prospective.

Spectral band selection is a fundamental problem in hyperspectral classification. This paper addresses the problem of band selection for hyperspectral remote sensing image and SVM parameter optimization. First, we present a thorough experime- ntal study to show the superiority of the generalization capability of the support vector machine (SVM) approach in the hyperspec- tral classification of remote sensing image. Second, we propose an evolutionary classification system based on particle swarm optimization (PSO) to improve the generalization performance of the SVM classifier. For this purpose, we have optimized the SVM classifier design by searching for the best value of the parameters that tune its discriminant function, and upstream by looking for the best subset of features that feed the classifier. The experiments are conducted on the basis of AVIRIS 92AV3C dataset. The obtained results clearly confirm the superiority of the SVM approach as compared to traditional classifiers, and suggest that further substantial improvements in terms of classification accuracy can be achieved by the proposed PSO- SVM classification system. important for a specific task. Some of them may be redundant or even irrelevant. Better performance may be achieved by discarding some features. In other circumstances, the dimensionality of input space may be decreased to save some computation effort, although this may slightly lower classification accuracy. Therefore, the classification process must be fast and accurate, using the smallest number of features. This objective can be achieved using feature selection. Feature selection strategies are often implied to explore the effect of irrelevant attributes on the performance of classifier systems. This study attempts to increase the classification accuracy rate by employing an approach based on particle swarm optimization (PSO) in SVM. This novel approach is termed PSO-SVM. The developed PSO-SVM approach not only tunes the parameter values of SVM, but also identifies a subset of features for specific problems, maximizing the classification accuracy rate of SVM. This makes the optimal separating hyper-plane obtainable in both linear and non-linear classification problems. In particular, they are organized so as to test the sensitivity of the SVM classifier and that of three reference classifiers used for comparison, i.e., SVM-Linear classifier, the k-nearest neighbor (K-nn) classifier and the radial basis function neural network (RBF-NN) classifier, with respect to the curse of dimensionality and the number of available training data

The variation in the distributions of recorded data between individuals leads to low classification accuracy. To address this issue, we introduce a multimodal fusion convolutional neural network (MFCNN). This network extracts common information from surface electromyography (sEMG) and accelerometer signals of different subjects using a two-stream convolutional neural network (CNN). To enhance the classification accuracy of a particular subject, a fine-tuning approach was implemented. The performance of the proposed method was assessed in four different scenarios, which include intersubject classification, intersubject classification when training data from multiple subjects, fine-tuned intersubject classification, and fine-tuned intersubject classification when training data from multiple subjects. The results demonstrate that in the intersubject scenario, when multiple subjects are available for training, the MFCNN achieves higher classification accuracy ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${p} &lt; 0.05$ </tex-math></inline-formula> ) than other neural networks and support vector machines (SVMs) that use sEMG signals [neural network (NN) and SVM], accelerometer signals (accNN and accSVM), sEMG and accelerometer signals [multimodal fusion nerual network (MFNN) and multimodal fusion support vector machine (MFSVM)] as inputs, as well as a CNN that uses sEMG signals as input after fine-tuning. Furthermore, compared with an MFCNN model trained with data from a single subject and an accCNN model trained with data from a single subject or multiple subjects, an MFCNN trained with multiple subjects demonstrated better performance on new subjects after fine-tuning ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${p} &lt; 0.05$ </tex-math></inline-formula> ). This method can learn common features among different subjects and improve the performance of classification among subjects. Our proposed method demonstrates the innovation of using a multimodal fusion approach and two-stream CNN to improve intersubject classification accuracy in upper limb movements.

Surface electromyogram (sEMG) signals are widely used to control the myoelectric prosthetic arm for amputees. In this study, the authors investigated the usefulness of discrete wavelet transform (DWT) features from multiple levels of approximation and detail coefficients obtained from sEMG signals. DWT is used for de-noising as well as feature extraction in this study and further tested using Support Vector Machine (SVM) classifier and Artificial Neural Network (ANN) Classifier. The performance of SVM classifier is compared with ANN classifier. The classification accuracy of SVM classifiers was found better as compared to ANN in term of speed and robustness. In the first section of paper the authors presented the introduction related to the study. The experimental set-up for recording surface EMG signal is presented in section 2. In section 3 the experimental results are depicted and in section 4 the results are concluded.

Aimed to the problem that it is hardship to get real-time and on-line measuring parameters in wood drying process, a novel PSO-SVM model that hybridized the particle swarm optimization (PSO) and support vector machines (SVM) to improve the nonlinearity caused by ambient temperature and other disturbance factors is presented. Support vector machines (SVM) based on statistical learning theory and structural risk minimization is proposed to deal with these problems. However, the model complexity and generalization performance of support vector machines (SVM) depend on a good setting of the three parameters (e,c,γ). In this paper, the particle swarm optimization is applied to optimize the parameters (e,c,γ) at the same time. Based on the proposed method, both PSO-SVM and SVM models are established and implemented to estimate lumber moisture content value in wood drying process. The result of comparative analysis is given. Experimental results show that solutions obtained by PSO-SVM training seem to be more robust and better generalization performance compared to SVM training.

The musculoskeletal disorder of a patient can be analyzed by using surface electromyogram (sEMG) signals. Its diagnosis is possible by classification of physical actions are bowing, clapping, handshaking, hugging, jumping, running, standing, seating, walking, and waving of surface-EMG signals. In this paper, an efficient method based on variational mode decomposition (VMD) is proposed for identification of physical activities of sEMG signals. VMD is an adaptive and non - recursive signal decomposition method which decomposes sEMG signals into several modes. These modes are used for extraction of statistical features like coefficient of variation, entropy, mean, negentropy, standard deviation, and zero crossing rate. Extracted features are fed into the multiclass least squares support vector machine (MC-LS-SVM) classifier with radial basis function (RBF) in order to classify normal physical actions of surface-EMG signals. The performance of obtained results shows that the method used provides a better classification accuracy of 98.17% for physical actions of surface-EMG signals as compared to existing methods.

Comparison of machine learning methods in sEMG signal processing for shoulder motion recognition

The learning vector quantization (LVQ), back propagation neural network (BPNN), and support vector machine (SVM) models were established to recognize face orientations. A precision function (P) was proposed to compute each model's precision with confusion matrix. The aforehand models were improved by intelligent algorithms to become LVQ with K-fold cross validation (CV-LVQ) model, BPNN with GA (GA-BPNN) model, and SVM with particle swarm optimization (PSO-SVM) model. The kernel function in the PSO-SVM model was assumed to RBF kernel which had relatively weaker learning ability. Hence the PSO-SVM model was further improved with a hybrid kernel that was fused with the generalization performance of global kernel and the learning ability of local kernel. The further improved PSO-SVM (IPSO-SVM) model possessed a 1.63 to 9.25 percent higher precision than PSO-SVM model. There were no obvious differences in the average elapsed time (AET) between IPSO-SVM model and PSO-SVM model. The results show that IPSO-SVM model not only reaches an outstanding precision of 98.14%, but also was practicable for the recognition of face orientations.

Myoelectric control has a wide range of potential applications including the design of human-machine interfaces for assistive technologies and robotics (prostheses and orthoses) as well as powered exoskeletons. The current work focuses on the Extreme Gradient Boosting - one of the most popular pattern recognition strategies for decoding the information of surface electromyography (sEMG) signals to infer the underlying muscle movements. In the EMG signal based controller design, it is now quite well-established that the position invariant model is a vital aspect. Hence, it should be considered to capture the inevitable dynamic nature of the upper limb. To this end, we have performed the experiments on a dataset consisting of the sEMG signals collected from eleven subjects at five different upper limb positions. The proposed method relies on the pre-processing stage of feature extraction which converts sEMG signal into the correlated time-domain descriptors (cTDD) - a set of descriptive values in the Euclidean space, which helps to learn the gradient boosting classifier. As the EMG signal classification is a subject-specific problem, the classifier has been customized and fine-tuned using the Bayesian optimization method for each subject to get the best possible results. The experimental results have shown that the proposed approach has outperformed the other existing popular classifiers in terms of classification accuracy.

The present work was aimed to present a thorough experimental study that shows the superiority of the generalisation capability of the support vector machine (SVM) approach in the classification of electrocardiogram (ECG) signals. Feature extraction was done using principal component analysis (PCA). Further, a novel classification system based on particle swarm optimisation (PSO) was used to improve the generalisation performance of the SVM classifier. For this purpose, we have optimised the SVM classifier design by searching for the best value of the parameters that tune its discriminant function and upstream by looking for the best subset of features that feed the classifier. The obtained results clearly confirm the superiority of the SVM approach as compared to traditional classifiers, and suggest that further substantial improvements in terms of classification accuracy can be achieved by the proposed PSO-SVM classification system.

Accurate Multi-class EMG signal classification is one of the key aspects of EMG-based prosthesis control. The other is a sufficient database. In this article, the process and classification of EMG signals are presented, which were recorded with the lightweight, easy-to-setup, semi-dry, 8-channeled, wireless MindRove Armband electrode system. Individual finger movements were captured with depth cameras, while the corresponding EMG signal was recorded. The labels about the executed movements were generated with a semi-automated algorithm. On the generated dataset Multiple classifiers, namely Random Forest, Extra Trees, Support Vector Machine, Nu-SVM, EEGNet, Ensemble, and Voting methods were tested and compared. Moreover, parameter searches were conducted, to increase the accuracy levels. In the case of EEGNet, the effect of transfer learning was also investigated.