Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models
Hand grasp recognition with surface electromyography (sEMG) has been used as a possible natural strategy to control hand prosthetics. However, effectively performing activities of daily living for users relies significantly on the long-term robustness of such recognition, which is still a challengin...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10017274/ |
| _version_ | 1797805034323312640 |
|---|---|
| author | Yuzhou Lin Ramaswamy Palaniappan Philippe De Wilde Ling Li |
| author_facet | Yuzhou Lin Ramaswamy Palaniappan Philippe De Wilde Ling Li |
| author_sort | Yuzhou Lin |
| collection | DOAJ |
| description | Hand grasp recognition with surface electromyography (sEMG) has been used as a possible natural strategy to control hand prosthetics. However, effectively performing activities of daily living for users relies significantly on the long-term robustness of such recognition, which is still a challenging task due to confused classes and several other variabilities. We hypothesise that this challenge can be addressed by introducing uncertainty-aware models because the rejection of uncertain movements has previously been demonstrated to improve the reliability of sEMG-based hand gesture recognition. With a particular focus on a very challenging benchmark dataset (NinaPro Database 6), we propose a novel end-to-end uncertainty-aware model, an evidential convolutional neural network (ECNN), which can generate multidimensional uncertainties, including vacuity and dissonance, for robust long-term hand grasp recognition. To avoid heuristically determining the optimal rejection threshold, we examine the performance of misclassification detection in the validation set. Extensive comparisons of accuracy under the non-rejection and rejection scheme are conducted when classifying 8 hand grasps (including rest) over 8 subjects across proposed models. The proposed ECNN is shown to improve recognition performance, achieving an accuracy of 51.44% without the rejection option and 83.51% under the rejection scheme with multidimensional uncertainties, significantly improving the current state-of-the-art (SoA) by 3.71% and 13.88%, respectively. Furthermore, its overall rejection-capable recognition accuracy remains stable with only a small accuracy degradation after the last data acquisition over 3 days. These results show the potential design of a reliable classifier that yields accurate and robust recognition performance. |
| first_indexed | 2024-03-13T05:45:59Z |
| format | Article |
| id | doaj.art-9ee955f0baad41d9917eb554bab5218d |
| institution | Directory Open Access Journal |
| issn | 1558-0210 |
| language | English |
| last_indexed | 2024-03-13T05:45:59Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| spelling | doaj.art-9ee955f0baad41d9917eb554bab5218d2023-06-13T20:10:25ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1558-02102023-01-013196297110.1109/TNSRE.2023.323698210017274Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware ModelsYuzhou Lin0https://orcid.org/0000-0003-3184-0523Ramaswamy Palaniappan1https://orcid.org/0000-0001-5296-8396Philippe De Wilde2https://orcid.org/0000-0002-4332-1715Ling Li3https://orcid.org/0000-0002-4026-0216School of Computing, University of Kent, Canterbury, Kent, U.K.School of Computing, University of Kent, Canterbury, Kent, U.K.Division of Natural Sciences, University of Kent, Canterbury, Kent, U.K.Department of Engineering, School of Science and Technology, City, University of London, London, Northampton Square, U.K.Hand grasp recognition with surface electromyography (sEMG) has been used as a possible natural strategy to control hand prosthetics. However, effectively performing activities of daily living for users relies significantly on the long-term robustness of such recognition, which is still a challenging task due to confused classes and several other variabilities. We hypothesise that this challenge can be addressed by introducing uncertainty-aware models because the rejection of uncertain movements has previously been demonstrated to improve the reliability of sEMG-based hand gesture recognition. With a particular focus on a very challenging benchmark dataset (NinaPro Database 6), we propose a novel end-to-end uncertainty-aware model, an evidential convolutional neural network (ECNN), which can generate multidimensional uncertainties, including vacuity and dissonance, for robust long-term hand grasp recognition. To avoid heuristically determining the optimal rejection threshold, we examine the performance of misclassification detection in the validation set. Extensive comparisons of accuracy under the non-rejection and rejection scheme are conducted when classifying 8 hand grasps (including rest) over 8 subjects across proposed models. The proposed ECNN is shown to improve recognition performance, achieving an accuracy of 51.44% without the rejection option and 83.51% under the rejection scheme with multidimensional uncertainties, significantly improving the current state-of-the-art (SoA) by 3.71% and 13.88%, respectively. Furthermore, its overall rejection-capable recognition accuracy remains stable with only a small accuracy degradation after the last data acquisition over 3 days. These results show the potential design of a reliable classifier that yields accurate and robust recognition performance.https://ieeexplore.ieee.org/document/10017274/Hand gesture recognitionsurface electromyography (sEMG)convolutional neural networktemporal variabilityrobustnessuncertainty |
| spellingShingle | Yuzhou Lin Ramaswamy Palaniappan Philippe De Wilde Ling Li Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models IEEE Transactions on Neural Systems and Rehabilitation Engineering Hand gesture recognition surface electromyography (sEMG) convolutional neural network temporal variability robustness uncertainty |
| title | Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models |
| title_full | Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models |
| title_fullStr | Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models |
| title_full_unstemmed | Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models |
| title_short | Robust Long-Term Hand Grasp Recognition With Raw Electromyographic Signals Using Multidimensional Uncertainty-Aware Models |
| title_sort | robust long term hand grasp recognition with raw electromyographic signals using multidimensional uncertainty aware models |
| topic | Hand gesture recognition surface electromyography (sEMG) convolutional neural network temporal variability robustness uncertainty |
| url | https://ieeexplore.ieee.org/document/10017274/ |
| work_keys_str_mv | AT yuzhoulin robustlongtermhandgrasprecognitionwithrawelectromyographicsignalsusingmultidimensionaluncertaintyawaremodels AT ramaswamypalaniappan robustlongtermhandgrasprecognitionwithrawelectromyographicsignalsusingmultidimensionaluncertaintyawaremodels AT philippedewilde robustlongtermhandgrasprecognitionwithrawelectromyographicsignalsusingmultidimensionaluncertaintyawaremodels AT lingli robustlongtermhandgrasprecognitionwithrawelectromyographicsignalsusingmultidimensionaluncertaintyawaremodels |