Development of classification methods for wheeze and crackle using mel frequency cepstral coefficient (MFCC): a deep learning approach
The most common method used by physicians and pulmonologists to evaluate the state of the lung is by listening to the acoustics of the patient's breathing by a stethoscope. Misdiagnosis and eventually, mistreatment are rampant if auscultation is not done properly. There have been efforts to add...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/86321/1/166-Article%20Text-1004-1-10-20201214.pdf http://irep.iium.edu.my/86321/ https://journals.iium.edu.my/kict/index.php/IJPCC https://doi.org/10.31436/ijpcc.v6i2.166 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The most common method used by physicians and pulmonologists to evaluate the state of the lung is by listening to the acoustics of the patient's breathing by a stethoscope. Misdiagnosis and eventually, mistreatment are rampant if auscultation is not done properly. There have been efforts to address this problem using a myriad of Machine Learning algorithms, but little has been done using Deep Learning. A Convolutional Neural Network (CNN) model with Mel Frequency Cepstral Coefficient (MFCC) is expected to mitigate these problems. The problem has been in the paucity of large enough datasets. Results show 0.76 and 0.60 for recall for wheeze and crackle respectively and these number are set to increase with optimization and larger, more diverse datasets. |
|---|