Classification Of EMG Signals Research Articles

Artificial intelligence for automatic classification of needle EMG signals: A scoping review

ObjectiveThis scoping review provides an overview of artificial intelligence (AI), including machine and deep learning techniques, in the interpretation of clinical needle electromyography (nEMG) signals. MethodsA comprehensive search of Medline, Embase and Web of Science was conducted to find peer-reviewed journal articles. All papers published after 2010 were included. The methodological quality of the included studies was assessed with CLAIM (checklist for artificial intelligence in medical imaging). Results51 studies were identified that fulfilled the inclusion criteria. 61% used open-source EMGlab data set to develop models to classify nEMG signal in healthy, amyotrophic lateral sclerosis (ALS) and myopathy (25 subjects). Only two articles developed models to classify signals recorded at rest. Most articles reported high performance accuracies, but many were subject to bias and overtraining. ConclusionsCurrent AI-models of nEMG signals are not sufficient for clinical implementation. Suggestions for future research include emphasizing the need for an optimal training and validation approach using large datasets of clinical nEMG data from a diverse patient population. SignificanceThe outcomes of this study and the suggestions made aim to contribute to developing AI-models that can effectively handle signal quality variability and are suitable for daily clinical practice in interpreting nEMG signals.

Design a 1D-CNN'S to classify surface EMG (SEMG) signals

Amputate forearm, finger, or hand is a biggest problem for the disabled subject. Therefore, Prosthetic does a significant role for the amputees to modify the capability and mobility of their systematic activities. Using the EMG signals of hand and finger motion discrimination are continuously growth for numerous hand and finger gestures. The main problem in designing a prosthetic hand is the classification of EMG signals. Machine learning (ML) algorithms present a solution to this problem by providing a way to classify EMG signals with simply and less costly scheme. This study presents more than one experiment on two datasets in order to classify individual fingers (IF) with wrist and victory based on a normative dataset of EMG signals and Deep Learning DL. These experiments show that the overall performance (average accuracy) of the proposed method is 98.83% and the overall error classification rate (error rate) is 1.17%.

MULTIFRACTAL ANALYSIS OF EMG FOR CLASSIFICATION AND PROGRESSIVE ASSESSMENT OF BICEPS BRACHII MUSCLE STRENGTH

The proposed research demonstrates an attempt to introduce a systematic procedure for studying the multifractal dynamics of biceps brachii muscle actions during light exercises. The intrinsic patterns in the Surface Electromyogram (sEMG) signals were extracted by fruitfully exploiting the Multifractal features of the signal. The Multifractal features are derived from the multifractal singularity spectrum of the EMG signals. This multifractal feature vector could be utilized for signal characterization, which was successfully extended for the classification of EMG signals. Experimental verification has been done to validate the feature extraction and classification algorithm proposed in this article. A pilot study was conducted on signals from the Physionet database, which was then extended to a medium database developed with biceps brachii EMG signals of 32 healthy male subjects. From this study, we could validate that the Multifractal features fit as a differentiating multi-feature set and also for the progressive assessment of EMG signals of different classes. The observations of the proposed method revealed that the strength of multifractality and area under the spectrum increase as a result of fast movements of the forearm or increases in muscle fatigue. The classification is performed using well-recognized supervised classification algorithms such as k-Nearest Neighbor and Support Vector Machine (SVM) Classifiers. The performance analysis of the classifiers are studied on various measures such as Accuracy, Precision, Recall, F1 score. The statistical significance analysis of the feature vector was carried out by one-way ANOVA test.

Performance improvement and complexity reduction in the classification of EMG signals with mRMR-based CNN-KNN combined model

For real-time applications, the performance in classifying the movements should be as high as possible, and the computational complexity should be low. This paper focuses on the classification of five upper arm movements which can be provided as a control for human-machine interface (HMI) based applications. The conventional machine learning algorithms are used for classification with both time and frequency domain features, and k-nearest neighbor (KNN) outplay others. To further improve the classification accuracy, pretrained CNN architectures are employed which leads to computational complexity and memory requirements. To overcome this, the deep convolutional neural network (CNN) model is introduced with three convolutional layers. To further improve the performance which is the key idea behind real-time applications, a hybrid CNN-KNN model is proposed. Even though the performance is high, the computation costs of the hybrid method are more. Minimum redundancy maximum relevance (mRMR), a feature selection method makes an effort to reduce feature dimensions. As a result, better performance is achieved by our proposed method CNN-KNN with mRMR which reduces computational complexity and memory requirement with a mean prediction accuracy of about 99.05±0.25% with 100 features.

Improving robotic hand control via adaptive Fuzzy-PI controller using classification of EMG signals

Robotic or prosthetic organs are designed to have the maximum similarity to human organs. This paper aims to improve robotic hand control via an adaptive Fuzzy-PI controller using EMG signals. The data is collected from the FDS and FPL muscles of the forearm of five individuals who performed eight movements. Then, appropriate filters are used to eliminate the noise of the signals, and MAV, VAR, and SE features are extracted. Based on MAV and VAR, classification is carried out using DA, KNN, and SVM. With an average accuracy, specificity, and sensitivity of 90.69%, 94.64%, and 62.10%, SVM is a better choice for movement detection. Following the movement detection by SVM, an appropriate reference signal is sent to the controller. The reference signal is the angle change of the fingers during the movement. All the eight gestures are modeled in a new way through these angles. The adaptive fuzzy-PI controller is used to control a robotic hand model with fifteen degrees of freedom. It has the advantages of learning from human experiences and adapting to environmental changes. The performance of the controller is evaluated in two ways. One is the comparison of the fuzzy-PI with the PI by RMSE. The average RMSE for eight movements using the fuzzy-PI is 1.6067, and for the PI, 5.0082. These results show that the fuzzy-PI controller performs better than the PI. Another new evaluation way presented in this paper is comparing the EMG signal features with the robotic hand movement signal features in terms of RMSE. The small RMSE values indicate that the EMG signal and robotic hand movement data features are significantly similar. Therefore, it can be concluded that the robotic hand controlled by the proposed controller is notably identical to the human hand.

Classification and signal processing analysis of the pathological electromyogram signal

This paper presents the classification of EMG signals for amyotrophic lateral sclerosis (ALS) disorder based on time-frequency method applying spectrogram. Four features [Shannon entropy (SSE), spectal entropy (SE), mean frequency (MNF) and median frequency (MDF)] are extracted from spectrogram in order to qualify the most appropriate one. So, two machine learning classifier of linear discriminate analysis (LDA) and support vector machine (SVM) are implemented to EMG signal. From the experiment, SVM classifier is outperformed others using SSE features from spectrogram which is more than 80% with an optimal window size of 512 ms. The finding of the study concludes that SVM is recommended to classify ALS disorder and can help rehabilitation centre to diagnose their patients in advanced state.

A Hybrid WD-EEMD sEMG Feature Extraction Technique for Lower Limb Activity Recognition

Classification and analysis of surface EMG (sEMG) signals have been of particular interest due to their numerous applications in the biomedical field. They can be used for the diagnosis of neuromuscular diseases, kinesiological studies, and human-machine interaction. However, these signals are difficult to process due to their noisy nature. To overcome this problem, a hybrid of wavelet with ensemble empirical mode decomposition pre-processing technique called WD-EEMD is proposed for classifying lower limb activities based on sEMG signals in healthy and knee abnormal subjects. First, Wavelet De-noising is used for filtering out white Gaussian Noise (WGN) and unwanted signals (contribution of other muscle signals). Next, an Ensemble Empirical Mode Decomposition is used for filtering out power line interference (PLI) and baseline wandering (BW) noises, followed by extraction of a total of nine time-domain features. Finally, the performance parameters of the Linear Discriminant Analysis (LDA) classifier are calculated with a 3-fold cross-validation technique. This study involves 11 healthy and 11 individuals with a knee abnormality for three different activities: walking, flexion of the leg up (standing), and leg extension from sitting position (sitting). Different pre-processing techniques similar to that of WD-EEMD were compared. It was observed that the proposed method achieves an average classification accuracy of 90.69% and 97.45% for healthy subjects and knee abnormal subjects, respectively.

A review on emg signal classification for human computer interaction

Human Computer Interaction (HCI) systems have become an increasingly essential aspect of our everyday lives as the importance of computerized devices in society grows. The efficient use of the available information flow of computer, communication, and display technologies is determined by HCI. In recent years, there has been a lot of buzz about developing intuitive interfaces that can detect and convert the user's body motions into computer instructions. Various biomedical signals (biosignals) that may be obtained from a specific tissue, organ, or cell system like the nervous system can be utilized for neural connection with computers. Electroencephalogram (EEG), Electrooculogram (EOG), and Electromyogram are some examples (EMG). Physically handicapped people benefit greatly from such methods. Many efforts have been made to create HCI using EMG signals from gesture. The development of continuous EMG signal categorization for graphical controller, which allows the physically handicapped to utilize word processing applications and other personal computer software, as well as the internet, is presently underway in the field of EMG signal processing and controller. It also allows for the manipulation of robotic devices, prosthetic limbs, I/O for virtual reality games, and fitness equipment, among other things. The majority of developmental research is focused on the use of neural networks to recognize patterns. The EMG controller may be configured to recognize gestures via signal analysis of muscle action potential groups. The aim of this review article is to examine the different methods and algorithms used for EMG signal categorization in order to convert EMG signals into computer commands.

Classification of Emg Signals Using Convolution Neural Network

An electrical signal is produced by the contraction of the muscles; this electrical signal contains information about the muscles, the recording of these signals called electromyography (EMG). This information is often used in studies such as prosthetic arm, muscle damage detection, and motion detection. Classifiers such as artificial neural networks, support vector machines are generally used for the classification of EMG signals. Despite successful results with such methods the extraction of the features to be given to the classifiers and the selection of the features affect the classification success. In this study, it is aimed to increase the success of the classification of the daily used hand movements using the Convolutional neural networks (CNN). The advantage of the deep learning techniques like CNN is that the relationships in big data are learned by the network. Firstly, the received EMG signals for forearms are windowed to increase the number of data and focus on the contraction points. Then, to compare the success rate, raw signals, Fourier transform of the signal, the root means square, and the Empirical mode decomposition (EMD) is applied to the signal and intrinsic mode functions are obtained. These signals are given to four different CNN. Afterward, to find the most efficient parameters, the results were obtained by splitting data set into three as 70% training set, 15% validation set, and 15% test set. 5 cross-validations have been applied to assess the system’s performance. The best results are obtained from the CNN, which receive the EMD applied signal as input. The result obtained with the cross-validation is 95.90% and the result obtained with the other separation method is 93.70%. When the results were examined, it was seen that CNN is a promising classifier even the raw signal is applied to the classifier. Also, it has been observed that EMD method creates better accuracy of classification.This is an open access article under the CC BY-SA 4.0 license.(https://creativecommons.org/licenses/by-sa/4.0/)

A Control System of SEMG Signal Based on Deep Learning

The research of control system based on sEMG signal is a popular field at present. It collects bioelectricity of human body through surface electrode. It has the new characteristic of subject fusion, and it is the combination of engineering technology and medical theory, specifically the application of cross combination of control science and electrophysiology. In this paper, the human surface EMG signal is taken as the research object, and a manipulator control system based on one-dimensional convolutional neural network (CNN) is proposed, and the functions and implementation methods of each part of the system are analyzed. The experimental results show that the recognition accuracy of the training model is 0.973, and the design scheme of EMG signal recognition and classification system with deep learning method is feasible. The successful design of the system provides technical support and theoretical basis for the further study of electrophysiological signals.

Neuromuscular Disease Detection Employing Deep Feature Extraction from Cross Spectrum Images of Electromyography Signals.

In this paper, a deep learning framework for detection and classification of EMG signals for diagnosis of neuromuscular disorders is proposed employing cross wavelet transform. Cross wavelet transform which is a modification of continuous wavelet transform is an important tool to analyze any non-stationary signal in time scale and in time-frequency frame. To this end, EMG signals of healthy, myopathy and Amyotrophic lateral sclerosis disorders were procured from an online existing database. A healthy EMG signal was chosen as reference and cross wavelet transform of the rest of the healthy as well as the disease EMG signals was done with the reference. From the resulting cross wavelet spectrum images of EMG signals, a convolution neural network (CNN) based automated deep feature extraction technique was implemented. The extracted deep features were further subjected to feature ranking employing one way analysis of variance (ANOVA) test. The extracted deep features with high degree of statistical significance were fed to several benchmark machine learning classifiers for the purpose of discrimination of EMG signals. Two binary classification problems are addressed in this paper and it has been observed that the highest mean classification accuracy of 100% is achieved using the statistically significant extracted deep features. The proposed method can be implemented for real-time detection of neuromuscular disorders.

Surface EMG signal classification using TQWT, Bagging and Boosting for hand movement recognition

Hands play a significant role in grasping and manipulating different objects. The loss of even a single hand have impact on the human activity. In this regard, a prosthetic hand is an appealing solution for the subjects who lost their hands. The surface electromyogram (sEMG) plays a vital role in the design of prosthesis hands. The ensemble classifiers achieve better performance by using a weighted combination of several classifier models. Hence, in this paper, the feasibility of the Bagging and the Boosting ensemble classifiers is assessed for the basic hand movement recognition by using sEMG signals, which were recorded during the grasping movements with various objects for the six hand motions. So, the novelty of the current study is the development of an ensemble model for hand movement recognition based on the tunable Q-factor wavelet transform (TQWT). The proposed method consists of three steps. In the first step, MSPCA is used for denoising. In the second step, a novel feature extraction method, TQWT is used for feature extraction from the sEMG signals, then, statistical values of TQWT sub-bands are calculated. In the last step, the obtained feature set is used as input to an ensemble classifier for the identification of intended hand movements. Performances of the Bagging and the Boosting ensemble classifiers are compared in terms of different performance measures. Using TQWT extracted features along with the presented the Adaboost with SVM and the Multiboost with SVM classifier results in a classification accuracy up to 100%. Hence, the results have shown that the proposed framework has achieved overall better performance and it is a potential candidate for the prosthetic hands control.

Surface EMG signal classification using ternary pattern and discrete wavelet transform based feature extraction for hand movement recognition

Hands are two of the most crucial organs and they play major role for human activities. Therefore, amputee people experience many difficulties in daily life. To overcome these difficulties, prosthetic hand is an effective solution. In order to automate the control of prosthetic hands, surface electromyogram (sEMG) signals and machine learning techniques play vital role. In this work, a novel ternary pattern and discrete wavelet (TP-DWT) based iterative feature extraction method is proposed. By using the proposed TP-DWT based feature extraction network, a sEMG signal recognition method is presented. The proposed TP-DWT based sEMG signal recognition method consists of channel concatenation, feature extraction using TP-DWT network, feature selection by using 2 leveled feature selection method and classification using conventional classifiers. The proposed method is tested by using a sEMG dataset, which were collected from amputee participants with 3 force levels (Low, Moderate, High). Four cases were studied to comprehensively evaluate the proposed TP-DWT based hand movements classification method with the sEMG signals. Based on the evaluations, the proposed TP-DWT based sEMG classification method achieved 99.14 % accuracy rate for all force levels by using k-nearest neighbor (k-NN) classifier with 10-fold cross validation. Moreover 97.78 %, 93.33 % and 92.96 % success rates are achieved for Low, Moderate and High force levels respectively.

Robust Classification of Intramuscular EMG Signals to Aid the Diagnosis of Neuromuscular Disorders.

Goal: This article presents the design and validation of an accurate automatic diagnostic system to classify intramuscular EMG (iEMG) signals into healthy, myopathy, or neuropathy categories to aid the diagnosis of neuromuscular diseases. Methods: First, an iEMG signal is decimated to produce a set of “disjoint” downsampled signals, which are decomposed by the lifting wavelet transform (LWT). The Higuchi's fractal dimensions (FDs) of LWT coefficients in the subbands are computed. The FDs of LWT subband coefficients are fused with one-dimensional local binary pattern derived from each downsampled signal. Next, a multilayer perceptron neural network (MLPNN) determines the class labels of downsampled signals. Finally, the sequence of class labels is fed to the Boyer-Moore majority vote (BMMV) algorithm, which assigns a class to every iEMG signal. Results: The MLPNN-BMMV classifier was experimented with 250 iEMG signals belonging to three categories. The performance of the classifier was validated in comparison with state-of-the-art approaches. The MLPNN-BMMV has resulted in impressive performance measures (%) using a 10-fold cross-validation—accuracy = n}{}99.87pm 0.25, sensitivity (normal) = n}{}99.97pm 0.13, sensitivity (myopathy) = n}{}99.68pm 0.95, sensitivity (neuropathy) = n}{}99.76pm 0.66, specificity (normal) = n}{}99.72pm 0.61, specificity (myopathy) = n}{}99.98pm 0.10, and specificity (neuropathy) = n}{}99.96pm 0.14—surpassing the existing approaches. Conclusions: A future research direction is to validate the classifier performance with diverse iEMG datasets, which would lead to the design of an affordable real-time expert system for neuromuscular disorder diagnosis.

Classification of EMG signal for multiple hand gestures based on neural network

<span>This paper presents the classification of EMG signal for multiple hand gestures based on neural network. In this study, the Electromyography is used to measure the muscle cell’s electrical activities which is commonly represented in a function time. Every muscle has their own signals, which was produced in every movement. Surface electromyography (sEMG) is used as a non-invasive technique for acquiring the EMG signal. The development of sensors’ detection and measuring the EMG have been improved and have become more precise while maintaining a small size. In this paper, the main objective is to identify the hand gestures based on: (1) Cylindrical Grasp, (2) Supination (Twist Left), (3) Pronation (Twist Right), (4) Resting Hand and (5) Open Hand that are predefined by using Arduino IDE, CoolTerm software and Microsoft Excel before using artificial neural network for classifying purposes in MATLAB. Finally, the extraction of the EMG patterns for each movement went through features extraction of the signals which is used to train the classifier in MATLAB to classify signals in the neural network. The features extracted are using mean absolute value (MAV), median, waveform length (WL) and root mean square (RMS). The Artificial Neural Network (ANN) produced accuracy of 80% for training and testing for 10 hidden neurons layer.</span>

An experimental study on upper limb position invariant EMG signal classification based on deep neural network

The classification of surface electromyography (sEMG) signal has an important usage in the man-machine interfaces for proper controlling of prosthetic devices with multiple degrees of freedom. The vital research aspects in this field mainly focus on data acquisition, pre-processing, feature extraction and classification along with their feasibility in practical scenarios regarding implementation and reliability. In this article, we have demonstrated a detailed empirical exploration on Deep Neural Network (DNN) based classification system for the upper limb position invariant myoelectric signal. The classification of eight different hand movements is performed using a fully connected feed-forward DNN model and also compared with the existing machine learning tools. In our analysis, we have used a dataset consisting of the sEMG signals collected from eleven subjects at five different upper limb positions. The time domain power spectral descriptors (TDPSD) is used as the feature set to train the DNN classifier. In contrast to the prior methods, the proposed approach excludes the feature dimensionality reduction step, which in turn significantly reduce the overall complexity. As the EMG signal classification is a subject-specific problem, the DNN model is customized for each subject separately to get the best possible results. Our experimental results in various analysis frameworks demonstrate that DNN based system can outperform the other existing classifiers such as k-Nearest Neighbour (kNN), Random Forest, and Decision Tree. The average accuracy obtained among the five subjects for DNN, SVM, kNN, Random Forest and Decision Tree is 98.88%, 98.66%, 90.64%, 91.78%, and 88.36% respectively. Moreover, it can achieve competitive performance with the state-of-the-art SVM based model, even though the proposed DNN model requires minimal processing in feature engineering. This study provides an insight into the detailed step-by-step empirical procedure to achieve the optimum results regarding classification accuracy using the DNN model.

Real‐time and user‐independent feature classification of forearm using EMG signals

AbstractElectromyography (EMG) signals contain various information about human motion. How to extract the EMG signals of the human body by appropriate methods for classification is a hot issue in current research. Unfortunately, the main problem with the classification of EMG signals is that only certain actions can be identified. Once the individual is changed, the recognition accuracy rate will be greatly reduced. This study introduces a method for classifying the forearm using back propagation (BP) neural networks. This mode extracted five features of the EMG signals. Participants were required to train their own actions during the test. Six participants selected four to six actions to identify them, and the average accuracy was more than 90%. The results suggest that the method can be used among different individuals and provides a good classification method.

Classification Method of Hand Gestures Based on Support Vector Machine

This paper presents the EMG signal classification based on PCA and SVM method. The data is acquired from the 5 subjects and each subject perform 7 hand gestures includes the tripod, power, precision closed, finger point, mouse, hand open, and hand close. Each gesture is repeated 10 times (5 data as training data and the 5 remaining data as testing data). Each of training and testing data are processed using 16 features extraction in time–domain and reduced using principal component analysis (PCA) to obtain new set of features. Features classification using support vector machine classify new set of features from each subject result 85% - 89% percentage of training classification. Training data classification is tested using testing data of EMG signals and giving accuracy reach 80% - 86%.

EMG finger movement classification based on ANFIS

An increase number of people suffering from stroke has impact to the rapid development of finger hand exoskeleton to enable an automatic physical therapy. Prior to the development of finger exoskeleton, a research topic yet important i.e. machine learning of finger gestures classification is conducted. This paper presents a study on EMG signal classification of 5 finger gestures as a preliminary study toward the finger exoskeleton design and development in Indonesia. The EMG signals of 5 finger gestures were acquired using Myo EMG sensor. The EMG signal features were extracted and reduced using PCA. The ANFIS based learning is used to classify reduced features of 5 finger gestures. The result shows that the classification of finger gestures is less than the classification of 7 hand gestures.

A Simple Network to Remove Interference in Surface EMG Signal from Single Gene Affected Phenylketonuria Patients for Proper Diagnosis

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.