MS-CLSTM: Myoelectric Manipulator Gesture Recognition Based on Multi-Scale Feature Fusion CNN-LSTM Network

Abstract

Surface electromyography (sEMG) signals reflect the local electrical activity of muscle fibers and the synergistic action of the overall muscle group, making them useful for gesture control of myoelectric manipulators. In recent years, deep learning methods have increasingly been applied to sEMG gesture recognition due to their powerful automatic feature extraction capabilities. sEMG signals contain rich local details and global patterns, but single-scale convolutional networks are limited in their ability to capture both comprehensively, which restricts model performance. This paper proposes a deep learning model based on multi-scale feature fusion-MS-CLSTM (MS Block-ResCBAM-Bi-LSTM). The MS Block extracts local details, global patterns, and inter-channel correlations in sEMG signals using convolutional kernels of different scales. The ResCBAM, which integrates CBAM and Simple-ResNet, enhances attention to key gesture information while alleviating overfitting issues common in small-sample datasets. Experimental results demonstrate that the MS-CLSTM model achieves recognition accuracies of 86.66% and 83.27% on the Ninapro DB2 and DB4 datasets, respectively, and the accuracy can reach 89% in real-time myoelectric manipulator gesture prediction experiments. The proposed model exhibits superior performance in sEMG gesture recognition tasks, offering an effective solution for applications in prosthetic hand control, robotic control, and other human-computer interaction fields.

Keywords: deep learning; multi-scale feature fusion; real-time prediction; sEMG gesture recognition.

Figure 1

Real-time control system for myoelectric manipulators based on sEMG. (a) sEMG data acquisition module; (b) sEMG processing and gesture recognition module; (c) Executive agency (Virtual manipulator).

Figure 2

sEMG acquisition: (a) Air-Band myoelectric bracelet; (b) Gesture action with Air-Band myoelectric bracelet; (c) sEMG acquisition.

Figure 3

Six hand movements: (1) index finger movement, (2) three-finger grip moving up and down, (3) left and right slide of the thumb, (4) thumbs up, (5) two-finger grasping and releasing, (6) clenched fist.

Figure 4

Grayscale diagram of sEMG. (Different colored lines in the figure represent data from 12 channels.)

Figure 5

The overall framework of the MS-CLSTM.

Figure 6

The structure of Multi-Scale Block.

Figure 7

The network structure of the CBAM.

Figure 8

Overall accuracy box plots for the MS-CLSTM.

Figure 9

Experimental scene diagram.

Figure 10

The recognition rate of 5 subjects based on MS-CLSTM.

Figure 11

Real-Time Recognition Accuracy Confusion Matrix. (Darker blue indicates higher accuracy, with a larger number of correctly classified instances. Lighter blue indicates lower accuracy, with fewer correctly classified instances.)