Search

This study delves into decoding hand gestures using surface electromyography (EMG) signals collected via a precision Myo-armband sensor, leveraging machine learning algorithms. The research entails rigorous data preprocessing to extract features and labels from raw EMG data. Following partitioning into training and testing sets, four traditional machine learning models are scrutinized for their efficacy in classifying finger movements across seven distinct gestures. The analysis includes meticulous parameter optimization and five-fold cross-validation to evaluate model performance. Among the models assessed, the Random Forest emerges as the top performer, consistently delivering superior precision, recall, and F1-score values across gesture classes, with ROC-AUC scores surpassing 99%. These findings underscore the Random Forest model as the optimal classifier for our EMG dataset, promising significant advancements in healthcare rehabilitation engineering and enhancing human-computer interaction technologies.

In this paper, a customizable wearable 3D-printed bionic arm is designed, fabricated, and optimized for a right arm amputee. An experimental test has been conducted for the user, where control of the artificial bionic hand is accomplished successfully using surface electromyography (sEMG) signals acquired by a multi-channel wearable armband. The 3D-printed bionic arm was designed for the low cost of 295 USD, and was lightweight at 428 g. To facilitate a generic control of the bionic arm, sEMG data were collected for a set of gestures (fist, spread fingers, wave-in, wave-out) from a wide range of participants. The collected data were processed and features related to the gestures were extracted for the purpose of training a classifier. In this study, several classifiers based on neural networks, support vector machine, and decision trees were constructed, trained, and statistically compared. The support vector machine classifier was found to exhibit an 89.93% success rate. Real-time testing of the bionic arm with the optimum classifier is demonstrated.

Introduction: The most common way of assessing surgical performance is by expert raters to view a surgical task and rate a trainee's performance. However, there is huge potential for automated skill assessment and workflow analysis using modern technology. The aim of the present study was to evaluate machine learning (ML) algorithms using the data of a Myo armband as a sensor device for skills level assessment and phase detection in laparoscopic training., Materials and Methods: Participants of three experience levels in laparoscopy performed a suturing and knot tying task on silicon models. Experts rated performance using Objective Structured Assessment of Surgical Skills (OSATS). Participants wore Myo armbands (Thalmic Labs™, Ontario, Canada) to record acceleration, angular velocity, orientation, and Euler orientation. ML algorithms (decision forest, neural networks, boosted decision tree) were compared for skill level assessment and phase detection., Results: 28 participants (8 beginner, 10 intermediate, 10 expert) were included, and 99 knots were available for analysis. A neural network regression model had the lowest mean absolute error in predicting OSATS score (3.7 ± 0.6 points, r 2 = 0.03 ± 0.81; OSATS min.-max.: 4-37 points). An ensemble of binary-class neural networks yielded the highest accuracy in predicting skill level (beginners: 82.2% correctly identified, intermediate: 3.0%, experts: 79.5%) whereas standard statistical analysis failed to discriminate between skill levels. Phase detection on raw data showed the best results with a multi-class decision jungle (average 16% correctly identified), but improved to 43% average accuracy with two-class boosted decision trees after Dynamic time warping (DTW) application., Conclusion: Modern machine learning algorithms aid in interpreting complex surgical motion data, even when standard analysis fails. Dynamic time warping offers the potential to process and compare surgical motion data in order to allow automated surgical workflow detection. However, further research is needed to interpret and standardize available data and improve sensor accuracy.

In this study, we present a transfer learning method for gesture classification via an inductive and supervised transductive approach with an electromyographic dataset gathered via the Myo armband. A ternary gesture classification problem is presented by states of 'thumbs up', 'thumbs down', and 'relax' in order to communicate in the affirmative or negative in a non-verbal fashion to a machine. Of the nine statistical learning paradigms benchmarked over 10-fold cross validation (with three methods of feature selection), an ensemble of Random Forest and Support Vector Machine through voting achieves the best score of 91.74% with a rule-based feature selection method. When new subjects are considered, this machine learning approach fails to generalise new data, and thus the processes of Inductive and Supervised Transductive Transfer Learning are introduced with a short calibration exercise (15 s). Failure of generalisation shows that 5 s of data per-class is the strongest for classification (versus one through seven seconds) with only an accuracy of 55%, but when a short 5 s per class calibration task is introduced via the suggested transfer method, a Random Forest can then classify unseen data from the calibrated subject at an accuracy of around 97%, outperforming the 83% accuracy boasted by the proprietary Myo system. Finally, a preliminary application is presented through social interaction with a humanoid Pepper robot, where the use of our approach and a most-common-class metaclassifier achieves 100% accuracy for all trials of a '20 Questions' game. [ABSTRACT FROM AUTHOR]

The effect of user experience on user satisfaction metrics in using MYO armband (hand gesture navigation) to control interactive maps and various other applications is still in a research state. In general, usability can be measured in terms of effectiveness, efficiency and satisfaction based on the metrics-task completion, error counts, task times and satisfaction scores. In this paper, we considered only satisfaction metrics. A simple and widely used System Usability Scale (SUS) questionnaire model is implemented to suggest some guidelines about the use of MYO armband. Another questionnaire, with a focus on ergonomic issues related to the use of the device such as Social Acceptability, Ease of use and ease of learning, Comfort and Stress, attempted do discover characteristics of hand gesture navigation using MYO. The results of this study can be used in a way to support the development of interactive maps using hand gesture navigation. [ABSTRACT FROM AUTHOR]

• Finger gestures are difficult to process due to overlapping layered structure of the arm muscles. • By examining the effects of gyroscope data, a new feature has been tried to be given to the literature. • EMG and gyroscope signals were acquired using Myo Armband with surface electrodes. • The data set used was shared in the supplementary documents section of the journal. • The highest performance features were determined using sequential forward feature selection. Gestures of the human hand can be identified through processing of surface electromyography (sEMG) signals. The human hand can perform many gestures via manipulation of the fingers. With correct classification of finger gestures, the mobility of a prosthetic hand can be increased and provide greater functionally. In this study, reliable classification was obtained for sEMG finger data acquired from a Myo armband placed on the lower forearm. In order to improve classification, gyroscopic signals, not previously used in other studies, were investigated in the sEMG finger data. Data was acquired from ten normal subjects using the Myo armband to identify 6 finger gestures: thumb, index finger, middle finger, little finger, ring finger and rest. Participants repeated each gesture thirty times. sEMG signals were preprocessed to extract features. 17 features were used in the feature matrix. By using the sequential forward feature selection method, the highest performance feature set was determined. Support Vector Machine, K-Nearest Neighbor and multilayer artificial neural network were used as classification algorithm. The classification was made using the Classification Learner Application and Neural Network Pattern Recognition Tool in Matlab®. The best performance with the features extracted only from sEMG data was 94.40% using the Artificial Neural Networks (ANN) method. The best performance with the features extracted from both sEMG and gyroscopic data was 96.30% (p-value < 0.05)with the ANN method. It is seen that gyroscopic signals can increase classification performance. [ABSTRACT FROM AUTHOR]