Feature recognition in multiple CNNs using sEMG images from a prototype comfort test

• Published in: Computer methods and programs in biomedicine Vol. 243; p. 107897
• Main Authors: Fu, You-Lei, Song, Wu, Xu, Wanni, Lin, Jie, Nian, Xuchao
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2024
• Subjects: Convolutional neural network; Health informatics; Prototype comfort; sEMG imaging; Sternocleidomastoid; MPF; EG; Prototype comfort; CG; Convolutional neural network; ROC; sEMG; sEMG imaging; SCM; Health informatics; Sternocleidomastoid
• ISSN: 0169-2607; 1872-7565; 1872-7565
• DOI: 10.1016/j.cmpb.2023.107897

•This study develops a methodology for combining sEMG with CNNs and applies it to the study of human-prototype interactions and the identification of body comfort.•Our innovative prototype usability experiment for sEMG measurements.•In this study, four CNNs, LeNet-5, VGGNet-11, inceptionnet V4 and DenseNet, were used to identify sEMG images.•Based on the comprehensive comparison of AUC results of the ROC curve, the best performance is VGGNet-11 (ACC = 0.92, AUC = 0.87).•The results of the study could provide a more suitable CNN model and serve as a useful tool to establish the comfort judgment of sEMG signals and deepen the development of products that are in close contact with the human body. Deep learning-based CNN networks have recently been investigated to solve the problem of body posture recognition based on surface electromyographic signals (sEMG). Influenced by these studies, to develop a combined approach of sEMG and CNNs in the study of human-product interactions and the impact of body comfort, and to compare the advantages and disadvantages of various CNNs networks. In this study, sEMG measurements were carried out by building a prototype usability experiment, and the data were divided into four categories, with two types of datasets: training and testing. Four CNNs, LeNet-5, VGGNet-11, InceptionNet V4, and DenseNet, were used for the recognition of sEMG images. DenseNet is another type of convolutional neural network with deep layers, which has a unique advantage over other algorithms. unique advantages over other algorithms. DenseNet has fewer layers and better accuracy than InceptionNet V4, but not only does it bypass enhanced feature reuse, but its network is easier to train and has some regularization effects, while also mitigating the problems of gradient disappearance and model degradation. These findings could lead to a more appropriate CNN model and a useful tool for developing comfort judgments of surface EMG signals, furthering the development of products that come into contact with the human body without the need for routine retraining.