Hand Motion Recognition from EMG using Artificial Neural Network (original) (raw)
Related papers
EMG (Electromyography) is a biological signal derived from the summation of electrical signals produced by muscular actions. This EMG can be integrated with external hardware and control prosthetics in rehabilitation. Pattern recognition plays an important role in developing myo-electric control based interfaces with prosthetics and Artificial Neural Networks (ANN) are widely used for such tasks. The main purpose of this paper is to classify different predefined hand gestured EMG signals (wrist up and finger flexion) using ANN and to compare the performances of four different Back propagation training algorithms used to train the network. The EMG patterns are extracted from the signals for each movement and then ANN is utilized to classify the EMG signals based on their features. The four different training algorithms used were SCG, LM, GD and GDM with different number of hidden layers and neurons. The ANNs were trained with those algorithms using the available experimental data as the training set. It was found that LM outperformed the others in giving the best performance within short time elapse. This classification can further be used to control devices based on EMG pattern recognition.
Design an System for Hand Gesture Recognition with Emg Signal by Neural Network
2022
The intellectual computing of an effective human-computer interaction (HCI) or human alternative and augmentative communication (HAAC) is vital in our lives in today's technological environment. One of the most essential approaches for developing a gesture-based interface system for HCI or HAAC applications is hand gesture recognition. As a result, in order to create an advanced hand gesture recognition system with successful applications, it is required to establish an appropriate gesture recognition technique. Human activity and gesture detection are crucial components of the rapidly expanding area of ambient intelligence, which includes applications such as robots, smart homes, assistive systems, virtual reality, and so on. We proposed a method for recognizing hand movements using surface electromyography based on an ANN . The CapgMyo dataset based on the Myo wristband (an eight-channel sEMG device) is utilized to assess participants' forearm sEMG signals in our technique...
Classification of EMG signal for multiple hand gestures based on neural network
Indonesian Journal of Electrical Engineering and Computer Science, 2020
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.
Neural Network Classifier for Hand Motion Detection from EMG Signal
2010
EMG signal based research is ongoing for the development of simple, robust, user friendly, efficient interfacing devices/systems for the disabled. The advancement can be observed in the area of robotic devices, prosthesis limb, exoskeleton, wearable computer, I/O for virtual reality games and physical exercise equipments. Additionally, electromyography (EMG) signals can also be applied in the field of human computer interaction (HCI) system. This paper represents the detection of different predefined hand motions (left, right, up and down) using artificial neural network (ANN). A backpropagation (BP) network with Levenberg-Marquardt training algorithm has been utilized for the classification of EMG signals. The conventional and most effective time and timefrequency based feature set is utilized for the training of neural network. The obtained results show that the designed network is able to recognize hand movements with satisfied classification efficiency in average of 88.4%. Furthermore, when the trained network tested on unknown data set, it successfully identify the movement types.
An implementation of movement classification for prosthesis control using custom-made EMG system
Serbian Journal of Electrical Engineering, 2017
Electromyography (EMG) is a well known technique used for recording electrical activity produced by human muscles. In the last few decades, EMG signals are used as a control input for prosthetic hands. There are several multifunctional myoelectric prosthetic hands for amputees on the market, but so forth, none of these devices permits the natural control of more than two degrees of freedom. In this paper we present our implementation of the pattern classification using custom made components (electrodes and an embedded EMG amplifier). The components were evaluated in offline and online tests, in able bodied as well as amputee subjects. This type of control is based on computing the time domain features of the EMG signals recorded from the forearm and using these features as input for a Linear Discriminant Analysis (LDA) classifier estimating the intention of the prosthetic user.
EMG Signal Processing for Hand Motion Pattern Recognition Using Machine Learning Algorithms
Archives of Orthopaedics
Electromyography (EMG) signal processing for assistive medical device control has been developed for clinical rehabilitation. The accuracy of operation and responsive time are still needed to be optimized. The purpose of this study was to determine and compare the efficiency of different artificial neural network-based machine learning (ML) algorithms in multiple channels surface EMG (sEMG) signal processing. EMG recorded from the forearm was processed for hand motion recognition. Performance of multilayer neural network training function "Trainlm" and "Trainscg" algorithms were evaluated based on their accuracy and duration required for EMG signal processing. The results showed both algorithms processed sEMG signal within less than 100 ms with the accuracy of Trainlm algorithm higher than Trainscg algorithm. The performance of the proposed methods was tested among five healthy subjects with accuracy higher than 98%. These outcomes suggested the Trainlm and Trainscg ML algorithms effectively recognized hand motion patterns, potentially they can be used for volitional control of hand robotic assistive device.
Development of Machine Learning Models to Determine Hand Gestures using EMG Signals
DAAAM Proceedings, 2020
The connection between the human body and science is still growing exponentially. The human body has many mysteries. The more we learn about them, the better we improve our scientific perspectives. In this case, the analysis of EMG (Electromyographic) signals gives the possibility the use the EMG data to perform classification tasks. Machine Learning models and Neural Networks are the best tools to classify different hand gestures using the dataset. This work aims to analyse the characteristics of EMG signals and use the EMG dataset to perform different ML models. The results will be used in robotic fields and control systems as future work. In this project, the Python programming language is used. The dataset was recorded using an MYO Thalmic bracelet. The number of instances is about 40000-50000 recordings in each column (channels). There are six different hand gestures tasks recorded in the dataset that are; hand clenched in a fist, wrist extension, wrist flexion, hand radial deviations, hand ulnar deviations, hand extended palm. The study of ML models and using gestures to control robotic devices could be useful in industrial spheres. A person may use a forearm bracelet to use it in industrial operations. Another purpose of this paper is; helping people who lost their hands. With the help of robotic arm plugged-in to their forearm and EMG device, they might be able to perform several hand-gestures.
LSTM Recurrent Neural Network for Hand Gesture Recognition Using EMG Signals
Applied Sciences, 2022
Currently, research on gesture recognition systems has been on the rise due to the capabili- ties these systems provide to the field of human–machine interaction, however, gesture recognition in prosthesis and orthesis has been carried out through the use of an extensive amount of channels and electrodes to acquire the EMG (Electromyography) signals, increasing the cost and complexity of these systems. The scientific literature shows different approaches related to gesture recognition based on the analysis of EMG signals using deep learning models, highlighting the recurrent neural networks with deep learning structures. This paper presents the implementation of a Recurrent Neural Network (RNN) model using Long-short Term Memory (LSTM) units and dense layers to develop a gesture classifier for hand prosthesis control, aiming to decrease the number of EMG channels and the overall model complexity, in order to increase its scalability for embedded systems. The proposed model requires the use of only four EMG channels to recognize five hand gestures, greatly reducing the number of electrodes compared to other approaches found in the literature. The proposed model was trained using a dataset for each gesture EMG signals, which were recorded for 20 s using a custom EMG armband. The model reached an accuracy of to 99% for the training and validation stages, and an accuracy of 87 ± 7% during real-time testing. The results obtained by the proposed model establish a general methodology for the reduction of complexity in the recognition of gestures intended for human.machine interaction for different computational devices.
2018 6th International Conference on Brain-Computer Interface (BCI)
Electromyography (EMG) signal analysis is one of the key determinants of the effectiveness of prosthetic devices. Modern researchers provide various methods of detection of different hand movements and postures. In this work, we examined the possibility to produce efficient detection of hand movement to a specific posture with the minimum possible number of electrodes. The data acquisition is produced with 1 channel BiTalino EMG sensor based on bipolar differential measurement. Using feature extraction and artificial neural network we achieved 82% of offline classification accuracy for 8 hand motions and 91% accuracy for 6 hand motions based on 200ms of EMG signal. Also, the motion detection algorithm was developed and successfully tested that allowed to implement the algorithm for real-time classification that showed sufficient accuracy for 2 and 4 motion classes cases.
Real-Time Hand Gesture Recognition Based on Artificial Feed-Forward Neural Networks and EMG
2018 26th European Signal Processing Conference (EUSIPCO), 2018
In this paper, we propose a real-time hand gesture recognition model. This model is based on both a shallow feedforward neural network with 3 layers and an electromyography (EMG) of the forearm. The structure of the proposed model is composed of 5 modules: data acquisition using the commercial device Myo armband and a sliding window approach, preprocessing, automatic feature extraction, classification, and postprocessing. The proposed model has an accuracy of 90.1% at recognizing 5 categories of gestures (fist, wave-in, wave-out, open, and pinch), and an average time response of 11 ms in a personal computer. The main contributions of this work include (1) a hand gesture recognition model that responds quickly and with relative good accuracy, (2) an automatic method for feature extraction from time series of varying length, and (3) the code and the dataset used for this work, which are made publicly available.