Big data in myoelectric control: large multi-user models enable robust zero-shot EMG-based discrete gesture recognition

被引:2
作者
Eddy, Ethan [1 ]
Campbell, Evan [1 ]
Bateman, Scott [2 ]
Scheme, Erik [1 ]
机构
[1] Univ New Brunswick, Inst Biomed Engn, Fredericton, NB, Canada
[2] Univ New Brunswick, Fac Comp Sci, Fredericton, NB, Canada
基金
加拿大自然科学与工程研究理事会;
关键词
big data; cross-user; deep learning; discrete; electromyography; gesture recognition; myoelectric control; zero-shot; PATTERN-RECOGNITION; TIME; CLASSIFICATION; EXTRACTION;
D O I
10.3389/fbioe.2024.1463377
中图分类号
Q81 [生物工程学(生物技术)]; Q93 [微生物学];
学科分类号
071005 ; 0836 ; 090102 ; 100705 ;
摘要
Myoelectric control, the use of electromyogram (EMG) signals generated during muscle contractions to control a system or device, is a promising input, enabling always-available control for emerging ubiquitous computing applications. However, its widespread use has historically been limited by the need for user-specific machine learning models because of behavioural and physiological differences between users. Leveraging the publicly available 612-user EMG-EPN612 dataset, this work dispels this notion, showing that true zero-shot cross-user myoelectric control is achievable without user-specific training. By taking a discrete approach to classification (i.e., recognizing the entire dynamic gesture as a single event), a classification accuracy of 93.0% for six gestures was achieved on a set of 306 unseen users, showing that big data approaches can enable robust cross-user myoelectric control. By organizing the results into a series of mini-studies, this work provides an in-depth analysis of discrete cross-user models to answer unknown questions and uncover new research directions. In particular, this work explores the number of participants required to build cross-user models, the impact of transfer learning for fine-tuning these models, and the effects of under-represented end-user demographics in the training data, among other issues. Additionally, in order to further evaluate the performance of the developed cross-user models, a completely new dataset was created (using the same recording device) that includes known covariate factors such as cross-day use and limb-position variability. The results show that the large data models can effectively generalize to new datasets and mitigate the impact of common confounding factors that have historically limited the adoption of EMG-based inputs.
引用
收藏
页数:24
相关论文
共 107 条
[1]   Electromyography data for non-invasive naturally-controlled robotic hand prostheses [J].
Atzori, Manfredo ;
Gijsberts, Arjan ;
Castellini, Claudio ;
Caputo, Barbara ;
Hager, Anne-Gabrielle Mittaz ;
Elsig, Simone ;
Giatsidis, Giorgio ;
Bassetto, Franco ;
Muller, Henning .
SCIENTIFIC DATA, 2014, 1
[2]   Characterization of a Benchmark Database for Myoelectric Movement Classification [J].
Atzori, Manfredo ;
Gijsberts, Arjan ;
Kuzborskij, Ilja ;
Elsig, Simone ;
Hager, Anne-Gabrielle Mittaz ;
Deriaz, Olivier ;
Castellini, Claudio ;
Mueller, Henning ;
Caputo, Barbara .
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, 2015, 23 (01) :73-83
[3]   The great time series classification bake off: a review and experimental evaluation of recent algorithmic advances [J].
Bagnall, Anthony ;
Lines, Jason ;
Bostrom, Aaron ;
Large, James ;
Keogh, Eamonn .
DATA MINING AND KNOWLEDGE DISCOVERY, 2017, 31 (03) :606-660
[4]  
Benalcazar Marco E, 2020, Zenodo
[5]  
Benalcázar ME, 2017, EUR SIGNAL PR CONF, P1040, DOI 10.23919/EUSIPCO.2017.8081366
[6]   Towards Open Set Deep Networks [J].
Bendale, Abhijit ;
Boult, Terrance E. .
2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, :1563-1572
[7]   Electromyography-Based Gesture Recognition: Is It Time to Change Focus From the Forearm to the Wrist? [J].
Botros, Fady S. ;
Phinyomark, Angkoon ;
Scheme, Erik J. .
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2022, 18 (01) :174-184
[8]   Context-informed incremental learning improves both the performance and resilience of myoelectric control [J].
Campbell, Evan ;
Eddy, Ethan ;
Bateman, Scott ;
Cote-Allard, Ulysse ;
Scheme, Erik .
JOURNAL OF NEUROENGINEERING AND REHABILITATION, 2024, 21 (01)
[9]   Deep Cross-User Models Reduce the Training Burden in Myoelectric Control [J].
Campbell, Evan ;
Phinyomark, Angkoon ;
Scheme, Erik .
FRONTIERS IN NEUROSCIENCE, 2021, 15
[10]   Current Trends and Confounding Factors in Myoelectric Control: Limb Position and Contraction Intensity [J].
Campbell, Evan ;
Phinyomark, Angkoon ;
Scheme, Erik .
SENSORS, 2020, 20 (06)