MoEMba: A Mamba-based Mixture of Experts for High-Density EMG-based Hand Gesture Recognition
#HD-sEMG #Hand Gesture Recognition #MoEMba #Selective State-Space Models #Channel Attention #Wavelet Feature Modulation #CapgMyo Dataset #Human-Computer Interaction
📌 Key Takeaways
- MoEMba framework improves HD-sEMG-based hand gesture recognition accuracy
- Achieves 56.9% balanced accuracy on CapgMyo dataset, outperforming previous methods
- Addresses 40% variability in inter-session and inter-subject classification
- Uses Selective State-Space Models with channel attention and wavelet feature modulation
- Shows potential for advancing HD-sEMG-powered HCI systems
📖 Full Retelling
Researchers Mehran Shabanpour, Kasra Rad, Sadaf Khademi, and Arash Mohammadi introduced the MoEMba framework, a novel approach for high-density electromyography (HD-sEMG) based hand gesture recognition, in their paper published on arXiv on February 24, 2026, aiming to address the significant challenge of low accuracy in inter-session and inter-subject classification that can reach up to 40% variability due to temporal inconsistencies in muscle signal measurements. The MoEMba framework represents an advancement in Human-Computer Interaction (HCI) technologies by leveraging Selective State-Space Models to enhance gesture recognition capabilities, specifically addressing temporal dependencies and cross-channel interactions through innovative channel attention techniques. Additionally, the researchers integrated wavelet feature modulation to capture multi-scale temporal and spatial relations, resulting in improved signal representation compared to previous methods. Experimental results demonstrated that MoEMba achieved a balanced accuracy of 56.9% on the CapgMyo HD-sEMG dataset, outperforming existing state-of-the-art approaches and showing significant potential for advancing HD-sEMG-powered HCI systems in practical applications.
🏷️ Themes
Human-Computer Interaction, Signal Processing, Machine Learning
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
--> Electrical Engineering and Systems Science > Signal Processing arXiv:2502.17457 [Submitted on 9 Feb 2025 ( v1 ), last revised 24 Feb 2026 (this version, v2)] Title: MoEMba: A Mamba-based Mixture of Experts for High-Density EMG-based Hand Gesture Recognition Authors: Mehran Shabanpour , Kasra Rad , Sadaf Khademi , Arash Mohammadi View a PDF of the paper titled MoEMba: A Mamba-based Mixture of Experts for High-Density EMG-based Hand Gesture Recognition, by Mehran Shabanpour and 3 other authors View PDF HTML Abstract: High-Density surface Electromyography has emerged as a pivotal resource for Human-Computer Interaction , offering direct insights into muscle activities and motion intentions. However, a significant challenge in practical implementations of HD-sEMG-based models is the low accuracy of inter-session and inter-subject classification. Variability between sessions can reach up to 40% due to the inherent temporal variability of HD-sEMG signals. Targeting this challenge, the paper introduces the MoEMba framework, a novel approach leveraging Selective StateSpace Models to enhance HD-sEMG-based gesture recognition. The MoEMba framework captures temporal dependencies and cross-channel interactions through channel attention techniques. Furthermore, wavelet feature modulation is integrated to capture multi-scale temporal and spatial relations, improving signal representation. Experimental results on the CapgMyo HD-sEMG dataset demonstrate that MoEMba achieves a balanced accuracy of 56.9%, outperforming its state-of-the-art counterparts. The proposed framework's robustness to session-to-session variability and its efficient handling of high-dimensional multivariate time series data highlight its potential for advancing HD-sEMG-powered HCI systems. Subjects: Signal Processing (eess.SP) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2502.17457 [eess.SP] (or arXiv:2502.17457v2 [eess.SP] for this version) https://doi.org/10.48550/arXiv.2502...
Read full article at source