Gesture Recognition Using MLP-Mixer With CNN and Stacking Ensemble for sEMG Signals

In recent years, gesture perception has become crucial to human–computer interaction (HCI) technologies. Among various techniques, gesture recognition based on surface electromyography (sEMG) signals has gained significant prominence, with deep-learning methods playing a pivotal role in this domain. However, as the demand for accurate gesture recognition continues to rise, there is a growing inclination toward selecting complex deep neural network architectures. This trend, however, poses challenges in terms of performance and runtime requirements for computing devices. This article introduces a novel gesture recognition method utilizing the multilayer perceptron (MLP)-Mixer framework combined with Stacking ensemble learning to address these challenges. The proposed method effectively captures the features of sEMG data by employing simple MLPs, achieving a level of accuracy comparable to complex networks while simultaneously reducing inference time. Experimental results demonstrate that the method performs a classification accuracy of 80.03% and 81.13% for 49 actions in the open-source dataset NinaPro DB2, using window lengths of 200 and 300 ms, respectively. Furthermore, the method achieves a single inference speed of 54.77 ms with a window length of 200 ms. In NinaPro DB5, with window lengths of 250 and 300 ms, the method presented in this article achieves accuracy rates of 73.39% and 74.82%, respectively, completing inference in just 11.45 ms using the 300-ms window length. Notably, the technique also demonstrates its ability to mitigate the impact of individual differences in sEMG data on recognition accuracy.