No abstract

The sensor patch system can be used for fitness guidance, skin disease detection, and wound monitoring and management by replacing related sensitive films.

No abstract

Muscular atrophy after limb fracture is a frequently occurring complication with multiple causes. Different treatments and targeted rehabilitation procedures should be carried out based on the causes. However, bedside evaluation methods are invasive in clinical practice nowadays, lacking reliable non-invasive methods. In this study, we propose a non-invasive flexible surface electromyography system with machine learning algorithms to distinguish nerve-injury and limb immobilization-related atrophy. First, a flexible surface electromyography sensor was designed and verified by in vitro tests for its robustness and flexibility. Then, in vivo tests on rats proved the reliability compared with the traditional invasive diagnosis method. Finally, this system was applied for the diagnosis of muscular atrophy in 10 patients. The flexible surface electromyography sensor can achieve a max strain of 12.0%, which ensures close contact with the skin. The in vivo tests on rats show great comparability with the traditional invasive diagnosis method. It can achieve a high specificity of 95.28% and sensitivity of 98.98%. Application on patients reaches a relatively high specificity of 89.44% and sensitivity of 91.94%. The proposed painless surface electromyography system can be an easy and accurate supplementary for bedside muscular atrophy causes evaluation, holding excellent contact with the body.

Surface electromyography (sEMG), as a key technology of non-invasive muscle computer interface, is an important method of human-computer interaction. We proposed a CNN-IndRNN (Convolutional Neural Network-Independent Recurrent Neural Network) hybrid algorithm to analyse sEMG signals and classify hand gestures. Ninapro’s dataset of 10 volunteers was used to develop the model, and by using only one time-domain feature (root mean square of sEMG), an average accuracy of 87.43% on 18 gestures is achieved. The proposed algorithm obtains a state-of-the-art classification performance with a significantly reduced model. In order to verify the robustness of the CNN-IndRNN model, a compact real¬time recognition system was constructed. The system was based on open-source hardware (OpenBCI) and a custom Python-based software. Results show that the 10-subject rock-paper-scissors gesture recognition accuracy reaches 99.1%.