胸部三轴加速度信号

The project used ECG patches provided by Vivalink, and data processing, data analysis, and algorithm construction were carried out after recruiting volunteers from Hangzhou University of Electronic Science and Technology (HUST) to complete the data collection of normal people, and the data collection of patients with pneumoconiosis from Shanghai Lung Hospital affiliated with Tongji University. The final work of cough recognition and distinguishing coughs of normal people and pneumoconiosis patients was achieved.Acceleration signals at the chest of healthy adults and patients with pneumoconiosis were collected using a triaxial acceleration sensor. Behaviours captured included: breathing, deep breathing, rapid breath change, coughing (in some experiments there was a specific subdivision: low intensity cough, medium intensity cough, high intensity cough, shut up cough), talking (in some experiments there was a specific subdivision: whispering, normal talking, loud talking), sighing, yawning. The acceleration signal was divided into 5s windows, ensuring that each 5s window must contain a certain behaviour. The acceleration signals were feature processed to obtain individual samples, each containing each computed feature value and its corresponding label.