Rejecting Novel Motions in High-Density Myoelectric Pattern Recognition Using Hybrid Neural Networks

Abstract

The objective of this study is to develop a method for alleviating a novel pattern interference toward achieving a robust myoelectric pattern-recognition control system. To this end, a framework was presented for surface electromyogram (sEMG) pattern classification and novelty detection using hybrid neural networks, i.e., a convolutional neural network (CNN) and autoencoder networks. In the framework, the CNN was first used to extract spatio-temporal information conveyed in the sEMG data recorded via high-density (HD) 2-dimensional electrode arrays. Given the target motion patterns well-characterized by the CNN, autoencoder networks were applied to learn variable correlation in the spatio-temporal information, where samples from any novel pattern appeared to be significantly different from those from target patterns. Therefore, it was straightforward to discriminate and then reject the novel motion interferences identified as untargeted and unlearned patterns. The performance of the proposed method was evaluated with HD-sEMG data recorded by two 8 × 6 electrode arrays placed over the forearm extensors and flexors of 9 subjects performing seven target motion tasks and six novel motion tasks. The proposed method achieved high accuracies over 95% for identifying and rejecting novel motion tasks, and it outperformed conventional methods with statistical significance (p < 0.05). The proposed method is demonstrated to be a promising solution for rejecting novel motion interferences, which are ubiquitous in myoelectric control. This study will enhance the robustness of the myoelectric control system against novelty interference.

Towards open-set text recognition via label-to-prototype learning

“…However, these extraordinary performances of classic closed-set systems are illusory, since their applications are limited when it comes to real and open world. Only a few studies [23], [24] have focused on rejecting novel gestures, but they overlook the importance of maintaining closed-set classification accuracy.…”

EMG-FRNet: A feature reconstruction network for EMG irrelevant gesture recognition

“…In this paper, the proposed EMG-FRNet for EMGbased irrelevant gesture recognition is compared with the existing methods based on SVDD (19) and AE (20). These comparative algorithms are all state-of-the-art methods in the field of irrelevant gesture recognition and have demonstrated good performance in this area, thus we chose them as the comparison algorithms in this study.…”

“…Therefore, the focus of this study lies in the recognition of irrelevant gestures. Existing methods in the field of irrelevant gesture recognition can be mainly categorized into probability-based approaches (13)(14)(15)(16)(17) and one-vs-all classification rule-based approaches (18)(19)(20)(21).…”

“…The purpose of SVDD is to find a minimum volume hyper-sphere to enclose samples belonging to the same target class. Wu et al (20)(21) used a set of classifiers composed of one-class autoencoder (AE) to address irrelevant motion interference. The purpose of AE is to reconstruct the input and judge whether the sample belongs to the target class based on the relationship between the reconstruction error and the threshold.…”