Gestures recognition based on surface electromyography (sEMG) has been widely used for human-computer interaction. However, there are few research studies on overcoming the influence of physiological factors among different individuals. In this paper, a cross-individual gesture recognition method based on long short-term memory (LSTM) networks is proposed, named cross-individual LSTM (CI-LSTM). CI-LSTM has a dual-network structure, including a gesture recognition module and an individual recognition module. By designing the loss function, the individual information recognition module assists the gesture recognition module to train, which tends to orthogonalize the gesture features and individual features to minimize the impact of individual information differences on gesture recognition. Through cross-individual gesture recognition experiments, it is verified that compared with other selected algorithm models, the recognition accuracy obtained by using the CI-LSTM model can be improved by an average of 9.15%. Compared with other models, CI-LSTM can overcome the influence of individual characteristics and complete the task of cross-individual hand gestures recognition. Based on the proposed model, online control of the prosthetic hand is realized.

中文翻译：

---

基于长短期记忆网络的跨个体手势识别

基于表面肌电图（sEMG）的手势识别已被广泛用于人机交互。然而，关于克服不同个体间生理因素影响的研究很少。本文提出了一种基于长短期记忆（LSTM）网络的跨个体手势识别方法，命名为跨个体LSTM（CI-LSTM）。CI-LSTM 具有双网络结构，包括手势识别模块和个体识别模块。通过设计损失函数，个体信息识别模块辅助手势识别模块进行训练，使手势特征和个体特征趋于正交，以尽量减少个体信息差异对手势识别的影响。通过跨个体手势识别实验，经验证，与其他选定的算法模型相比，使用CI-LSTM模型获得的识别准确率平均可提高9.15%。与其他模型相比，CI-LSTM 可以克服个体特征的影响，完成跨个体手势识别任务。基于所提出的模型，实现了假手的在线控制。