Respiratory sound 2-class classification on ICBHI 2017 (80/20)
0.833SpecificityRespireNet (CNN base)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| RespireNet (CNN base)architecture=CNN2020.10 | 0.833 | 0.605 | 0.719 | |
| RespireNet (CNN+CBA+BRC+FT)architecture=CNN, augmentation=Concatenation-based augmentation, clipping=Blank region clipping, fine-tuning=device specific2020.10 | 0.809 | 0.731 | 0.77 | |
| RespireNet (CNN+CBA+BRC)architecture=CNN, augmentation=Concatenation-based augmentation, clipping=Blank region clipping2020.10 | 0.764 | 0.71 | 0.737 |