Article
Chemistry, Analytical
Nebojsa Malesevic, Anders Bjorkman, Gert S. Andersson, Christian Cipriani, Christian Antfolk
Summary: This paper evaluates fourteen common algorithms for the direct and proportional control of a prosthetic hand. The estimation of forces generated in the hand using different algorithms is compared to the measured forces, providing a baseline performance metric for more advanced algorithms.
Article
Engineering, Electrical & Electronic
Hyeyun Lee, Soyoung Lee, Jaeseong Kim, Heesoo Jung, Kyung Jae Yoon, Srinivas Gandla, Hogun Park, Sunkook Kim
Summary: With the help of AI-based algorithms, the accuracy of gesture recognition using sEMG signals has increased. An array of bipolar stretchable sEMG electrodes, combined with a self-attention-based graph neural network, is developed to achieve highly accurate gesture recognition. The system can differentiate static and dynamic gestures with about 97% accuracy using a single trial per gesture. The array also has skin-like attributes and can provide stable EMG signals even after long-term testing and multiple reuses.
NPJ FLEXIBLE ELECTRONICS
(2023)
Article
Chemistry, Analytical
Francisco Perez-Reynoso, Nein Farrera-Vazquez, Cesar Capetillo, Nestor Mendez-Lozano, Carlos Gonzalez-Gutierrez, Emmanuel Lopez-Neri
Summary: This study presents principles for the development of interfaces in physiotherapy or rehabilitation assistance systems and proposes a solution to the customization problem. By utilizing sEMG database and neural networks, the system enables personalized therapy and adaptation to individuals. The results demonstrate that customizing the interface reduces the learning curve, enhancing the effectiveness of the system.
Article
Computer Science, Information Systems
Sujiao Li, Yue Zhang, Yuanmin Tang, Wei Li, Wanjing Sun, Hongliu Yu
Summary: Currently, sEMG-based pattern recognition is a crucial and promising control method for prosthetic limbs. A 1D convolutional recurrent neural network classification model is proposed to address the issue of simultaneous consideration of classification recognition rate and time delay. Offline experiments verify the recognition performance, while online experiments examine real-time recognition performance and time delay. The 1D-CNN-RNN classification model demonstrates higher performances in real-time recognition accuracy and shorter time delay, making it an efficient control for dexterous prostheses.
Article
Engineering, Biomedical
Xiaodong Liu, Enhao Zheng, Qining Wang
Summary: In this study, a high-framerate Electrical Impedance Tomography (EIT) system combined with an adaptive recognition algorithm was proposed for real-time wrist kinematics decoding. The EIT-based interface provided deep muscular spatial information and maintained consistent muscle morphology after sensor re-donning. The results demonstrated the potential of the EIT-based interface in real-time human motion intent recognition.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Neurosciences
Mojisola Grace Asogbon, Oluwarotimi Williams Samuel, Ejay Nsugbe, Yongcheng Li, Frank Kulwa, Deogratias Mzurikwao, Shixiong Chen, Guanglin Li
Summary: This article investigates the impact of myoelectric signal recording duration on finger gesture recognition and finds that a recording duration of 5 to 10 seconds can achieve good decoding accuracy. It also provides guidance for selecting appropriate recording durations.
FRONTIERS IN NEUROSCIENCE
(2023)
Article
Engineering, Biomedical
Chen Chen, Yang Yu, Xinjun Sheng, Dario Farina, Xiangyang Zhu
Summary: The study extends offline EMG decomposition algorithm to real-time identification of motor unit activities and proposes a MU-based method for online control of multiple motor tasks. Experimental results show good accuracy in identifying MUs and highly correlated activation of specific motions, demonstrating superior performance compared to conventional myo-control methods.
JOURNAL OF NEURAL ENGINEERING
(2021)
Article
Engineering, Multidisciplinary
Haiqiang Duan, Chenyun Dai, Wei Chen
Summary: The development of wearable smart bracelets has attracted more attention in the field of human-machine interfaces, but the limited data collection range requires studying the cooperation between wrist and finger movements to simplify the data collection system. This study quantified the HD-sEMG forearm spatial activation features and performed linear fitting to verify the linear superposition relationship between wrist and finger movements. Four commonly adopted classifiers were used to classify and predict the results of gesture fitting.
MATHEMATICAL PROBLEMS IN ENGINEERING
(2022)
Article
Chemistry, Multidisciplinary
Kai Way Li, Thi Lan Anh Nguyen
Summary: An experiment was conducted to investigate the movement time and subjective rating of difficulty for real and virtual pipe transferring tasks. It was found that the movement time for transferring a virtual pipe was significantly shorter than that for transferring a real pipe. Male participants transferred the pipe faster than their female counterparts. The movement time for both real and virtual object transferring was influenced by gender, handedness, and transferring direction. Additionally, the subjective rating of difficulty was positively correlated with the movement time.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Biomedical
Mohamed Z. Amrani, Christoph W. Borst, Nouara Achour
Summary: This study introduces and evaluates a multi-sensory hand pattern recognizer that utilizes MLP and DT classifiers to achieve optimal hand shape detection. By integrating LM controllers, VIVE Pro Eye camera, and MYO armband, the study successfully reaches an average classification accuracy of 98.31%.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Engineering, Biomedical
Dezhen Xiong, Daohui Zhang, Xingang Zhao, Yaqi Chu, Yiwen Zhao
Summary: This paper presents a novel approach using SPD manifold to extract spatial structural information from limited EMG signals for myoelectric pattern recognition. Experimental results demonstrate the superior performance of the proposed method compared to traditional feature extraction methods, and its effectiveness is further validated on different databases.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Yanting Li, Junwei Jin, C. L. Philip Chen
Summary: This paper proposes a new method for face recognition by incorporating unselected training samples into the modeling process, improving recognition accuracy and the effectiveness of classification.
JOURNAL OF REAL-TIME IMAGE PROCESSING
(2021)
Article
Psychology, Multidisciplinary
Thomas Jacquet, Benedicte Poulin-Charronnat, Patrick Bard, Romuald Lepers
Summary: The study found that subjective mental fatigue significantly increased after completing a cognitive task, but decreased during the 20-minute recovery period without returning to pre-task levels. EEG recordings during the recovery period showed an increase in theta and alpha power over time, indicating a persistence of mental fatigue. Arm-pointing movement duration gradually increased during the recovery period, suggesting that behavioral performance remained impaired 20 minutes after the cognitively demanding task.
FRONTIERS IN PSYCHOLOGY
(2021)
Article
Engineering, Biomedical
Xiangxin Li, Lan Tian, Yue Zheng, Oluwarotimi Williams Samuel, Peng Fang, Lin Wang, Guanglin Li
Summary: Surface electromyogram pattern recognition (EMG-PR) is a promising approach for predicting amputees' motion intentions to control myoelectric prostheses. In this study, a feature filtering strategy was proposed and applied to improve the performance of EMG-PR. Experimental results showed that the proposed strategy significantly increased motion classification accuracy and real-time motion completion rate.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Engineering, Multidisciplinary
Lizhi Pan, Kai Liu, Kun Zhu, Jianmin Li
Summary: Amputees have poorer performances in EMG pattern recognition compared to able-bodied individuals, and factors such as muscle weakness and atrophy, limb length, and motor cortex decrease have been studied. However, the impact of the absence of joint movements has not been explored. This study found that hand and wrist joint movements significantly affect EMG pattern recognition, providing a new perspective for future research.
JOURNAL OF BIONIC ENGINEERING
(2022)
Article
Computer Science, Information Systems
Kimia Dinashi, Ali Ameri, Mohammad Ali Akhaee, Kevin Englehart, Erik Scheme
Summary: This study proposes a new method for efficient compression of EMG data using deep convolutional autoencoders (CAE), achieving significant results in experiments. The CAE architecture can generate a highly compressed abstract data representation without significantly affecting the accuracy of data classification. Additionally, the method demonstrates excellent inter-subject performance and high generalizability.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2022)
Article
Chemistry, Analytical
Ines Chihi, Lilia Sidhom, Ernest Nlandu Kamavuako
Summary: This paper presents a novel approach for characterising muscle force using electromyography (EMG) signals. The proposed method, based on a nonlinear Hammerstein-Wiener model, effectively estimates muscle force with reasonable accuracy. Compared to traditional artificial neural network methods, this approach demonstrates higher accuracy and reliability. The findings suggest that the use of a multimodel approach is important for improving proportional control accuracy in prostheses.
Article
Computer Science, Artificial Intelligence
Muhammad Akmal, Syed Zubair, Mads Jochumsen, Muhammad Zia Ur Rehman, Ernest Nlandu Kamavuako, Muhammad Irfan Abid, Imran Khan Niazi
Summary: This paper focuses on the key issue in designing a prosthetic hand that can classify movements based on electromyography (EMG) signals. It explores the use of multiday intramuscular EMG signals and applies matrix and tensor factorization methods to recover missing data. The results show that CP-WOPT outperforms other methods in recovering large percentage of missing data and exhibits robustness.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2022)
Article
Chemistry, Analytical
Xinqi Bao, Yujia Xu, Ernest Nlandu Kamavuako
Summary: This study analyzed the effect of signal duration on the performance of deep learning methods for heart sound classification. The results showed that short signal duration weakened the performance of recurrent neural networks (RNNs), while it had little impact on convolutional neural networks (CNNs). RNNs outperformed CNNs when using Mel-frequency cepstrum coefficients (MFCCs) as features. The addition of dynamic information had minimal improvement on the performance of both RNNs and CNNs.
Article
Chemistry, Analytical
Carlotta Malvuccio, Ernest N. Kamavuako
Summary: As the elderly population increases, there is a need for healthcare technologies to monitor fluid intake. This study investigated the potential of combining optimal features from sEMG sensors to improve classification and estimation accuracy. Results showed promising accuracy in differentiating between liquid swallows and non-liquid swallowing events, as well as estimating the volume of fluid intake.
Editorial Material
Chemistry, Analytical
Ernest N. Kamavuako
Article
Engineering, Electrical & Electronic
Muhammad Farrukh Qureshi, Zohaib Mushtaq, Muhammad Zia Ur Rehman, Ernest Nlandu Kamavuako
Summary: This research compares multiday surface EMG recordings and measures the performance of a convolutional neural network (CNN) in enhancing myoelectric control. The proposed CNN achieves high accuracy in classifying EMG data from able-bodied individuals and amputees. It outperforms other classifiers in terms of accuracy and computational cost.
IEEE SENSORS JOURNAL
(2022)
Article
Chemistry, Analytical
Bingbin Wang, Levi Hargrove, Xinqi Bao, Ernest N. Kamavuako
Summary: This study compared the statistical properties of surface electromyography signals used in home and laboratory settings for prosthesis calibration, finding differences in between-calibration classification errors but not within-calibration classification errors.
Article
Engineering, Electrical & Electronic
Muhammad Farrukh Qureshi, Zohaib Mushtaq, Muhammad Zia Ur Rehman, Ernest Nlandu Kamavuako
Summary: In this study, an efficient concatenated convolutional neural network (E2CNN) is proposed for classification of surface electromyography (sEMG) extracted from the upper limb. The E2CNN model shows good response time and high accuracy when applied to LMS-based images. It has the potential to be a candidate model for real-time classification of sEMG based on LM spectrogram images.
IEEE SENSORS JOURNAL
(2023)
Article
Chemistry, Analytical
Iman Ismail, Imran Khan Niazi, Heidi Haavik, Ernest N. Kamavuako
Summary: This paper investigates the effect of sEMG features on the detection of drinking events and estimation of the amount of water swallowed per sip. The results show that sEMG can accurately distinguish drinking events from non-drinking events and provide a good estimation of fluid intake. These findings validate the potential and importance of sEMG in monitoring and estimating fluid intake.
Review
Chemistry, Analytical
Xin Chen, Ernest N. Kamavuako
Summary: Food and fluid intake monitoring is crucial for preventing dehydration, malnutrition, and obesity. While dietary monitoring has received more attention, fluid intake monitoring has been neglected. Vision-based methods offer non-intrusive solutions for monitoring, but face challenges in areas such as occlusion, privacy, computational efficiency, and practicality. This paper reviews existing research on vision-based intake monitoring methods, assesses the available literature, and identifies current challenges and research gaps.
Article
Biotechnology & Applied Microbiology
Lilia Sidhom, Ines Chihi, Mahfoudh Barhoumi, Nesrine Ben Afia, Ernest Nlandu Kamavuako, Mohamed Trabelsi
Summary: This paper aims to design a smart biosensor to predict ECG signals in a specific auscultation site, using a hybrid architecture of Artificial Neural Networks (ANNs) and Taguchi optimizer. By optimizing the input factors and improving the prediction accuracy, the developed biosensor outperforms the ANN-based biosensor in predicting ECG signals from different measurement sites.
BIOENGINEERING-BASEL
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Xinqi Bao, Fenghe Hu, Yujia Xu, Mohamed Trabelsi, Ernest Kamavuako
Summary: Paroxysmal atrial fibrillation is a type of atrial fibrillation that occurs rapidly and stops spontaneously. This study proposes a two-stage algorithm that can accurately classify and identify the onset of paroxysmal atrial fibrillation in real time.
BIOSIGNALS: PROCEEDINGS OF THE 15TH INTERNATIONAL JOINT CONFERENCE ON BIOMEDICAL ENGINEERING SYSTEMS AND TECHNOLOGIES - VOL 4: BIOSIGNALS
(2022)
Article
Engineering, Biomedical
Jena L. Nawfel, Kevin B. Englehart, Erik J. Scheme
Summary: Studies have shown that closed-loop myoelectric control schemes can impact user performance and behavior compared to open-loop systems. Visual feedback provided during user training can influence the quality and predictability of a myoelectric classification-based control system. The commonly used screen guided training protocol may not represent online use effectively, suggesting the need for better training protocols that mimic real-time control.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2022)
Proceedings Paper
Engineering, Biomedical
Carlotta Malvuccio, Ernest N. Kamavuako
Summary: Dehydration is a serious concern for older adults, and this study investigates the use of sEMG sensors to detect swallowing events and estimate fluid intake volume. The results show high accuracy in distinguishing between noise and swallows, but further validation is needed for volume estimation, especially for continuous swallows. This study suggests the promising potential of using surface EMGs to track fluid intake.
2020 IEEE-EMBS CONFERENCE ON BIOMEDICAL ENGINEERING AND SCIENCES (IECBES 2020): LEADING MODERN HEALTHCARE TECHNOLOGY ENHANCING WELLNESS
(2021)