Inter-Subject Domain Adaptation for CNN-Based Wrist Kinematics Estimation Using sEMG

• Published in: IEEE transactions on neural systems and rehabilitation engineering Vol. 29; pp. 1068 - 1078
• Main Authors: Bao, Tianzhe, Zaidi, Syed Ali Raza, Xie, Shengquan, Yang, Pengfei, Zhang, Zhi-Qiang
• Format: Journal Article
• Language: English
• Published: New York IEEE 2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptation; Adaptation models; Artificial neural networks; CNN; Divergence; domain adaptation; Domains; Electromyography; Estimation; Feature extraction; Kinematics; Neural networks; Regression; sEMG; Supervised learning; Training; Transfer learning; Wrist; wrist kinematics estimation
• ISSN: 1534-4320; 1558-0210; 1558-0210
• DOI: 10.1109/TNSRE.2021.3086401

Recently, convolutional neural network (CNN) has been widely investigated to decode human intentions using surface Electromyography (sEMG) signals. However, a pre-trained CNN model usually suffers from severe degradation when testing on a new individual, and this is mainly due to domain shift where characteristics of training and testing sEMG data differ substantially. To enhance inter-subject performances of CNN in the wrist kinematics estimation, we propose a novel regression scheme for supervised domain adaptation (SDA), based on which domain shift effects can be effectively reduced. Specifically, a two-stream CNN with shared weights is established to exploit source and target sEMG data simultaneously, such that domain-invariant features can be extracted. To tune CNN weights, both regression losses and a domain discrepancy loss are employed, where the former enable supervised learning and the latter minimizes distribution divergences between two domains. In this study, eight healthy subjects were recruited to perform wrist flexion-extension movements. Experiment results illustrated that the proposed regression SDA outperformed fine-tuning, a state-of-the-art transfer learning method, in both single-single and multiple-single scenarios of kinematics estimation. Unlike fine-tuning which suffers from catastrophic forgetting, regression SDA can maintain much better performances in original domains, which boosts the model reusability among multiple subjects.