2023 Fiscal Year Final Research Report
Development of automatic cough monitoring, measurement and service system
| Project/Area Number |
20K12080
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 62020:Web informatics and service informatics-related
|
|---|
| Research Institution | The University of Aizu |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
松井 知子 統計数理研究所, モデリング研究系, 教授 (10370090)
齋藤 純平 福島県立医科大学, 医学部, 講師 (50332929)
|
|---|
| Project Period (FY) |
2020-04-01 – 2024-03-31
|
| Keywords | cough monitoring / health informatics / audio event detection |
|---|
| Outline of Final Research Achievements |
In this research, we propose a novel deep learning system for the efficient detection and monitoring of cough events in audio recordings. First, we perform voice activity detection to eliminate audio silences. Next, we employ a cough classification to identify the presence of cough. Finally, we implement cough event detection using a high-performance classification-regression fusion method. Our approach is unique in several notable ways: 1) we incorporate a teacher-student framework for the training of our detection model, 2) instead of relying on specific audio features such as MFCC or Mel Spectrogram, our end-to-end system takes the raw audio signal directly as input and outputs the cough boundary timings, 3) the proposed method is general enough to be used for various other sound event monitoring tasks. The comparative cough monitoring evaluation of our approach against systems such as the Leicester Cough Monitor demonstrates our method's superiority by achieving the lowest error.
|
| Free Research Field |
audio event detection
|
| Academic Significance and Societal Importance of the Research Achievements |
The proposed cough detection and monitoring approach has the potential to impact various domains beyond respiratory health, such as urban planning, environmental monitoring, and animal welfare. The methodology is independent and adaptable to any sound event detection task requiring counting events.
|
