Respiratory sound signal quality estimation

• Main Author: Zhang, Ziqing
• Other Authors: Ser Wee
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/142412
• Institution: Nanyang Technological University

As a major disease leading to death in recent years, cardiopulmonary disease is an endemic disease that is difficult to treat in many parts of the world. Its diagnosis and treatment difficulties are mainly reflected in the low efficiency of the diagnosis and treatment method, the diagnosis and treatment accuracy and the discomfort brought to patients during the treatment. Given the technical limitations of the way the world currently treats such conditions, it is extremely important and urgent for the quality of signals to be controlled accurately. This study mainly utilizes MATLAB programming, the basic knowledge of machine learning and signal processing to complete estimation and examination of the quality of respiratory sound signals, including data collection, feature extraction, feature selection, classification and quality examination. All of the data came from real patients and healthy people, which guarantees the whole study conducted truly and effectively. In the process of feature analysis, necessary auxiliary tools, such as FR, MATLAB programming and SVM classifier, all help the research in an orderly manner. For the analysis and quality estimation of respiratory sound signal, the evaluation criterion is the final quality inspection accuracy. In this study, the two feature groups respectively determined by MFCC 3,2,1 and MFCC 3,2,1,8 indicating respectively good test accuracy, which are believed to be applicable to the diagnosis and treatment of respiratory diseases, mainly for the detection of respiratory sound signal quality.