Respiratory Sound Classification on SNUBH in-house clinical
88.78SpecificityMeta-Ensemble
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Meta-EnsembleSplit strategy=5-Fold, Data granularity level=P-level2026.04 | 88.78 | 69.5 | 79.14 | |
| Meta-EnsembleSplit strategy=5-Fold, Data granularity level=S-level2026.04 | 83.55 | 73 | 78.28 | |
| Meta-EnsembleSplit strategy=Fixed Split, Data granularity level=P-level2026.04 | 80.24 | 77.5 | 78.82 | |
| Kim et al. (BTS)Training protocol=BTS2026.04 | 77.38 | 76.14 | 76.76 | |
| Meta-EnsembleSplit strategy=Fixed Split, Data granularity level=S-level2026.04 | 76.4 | 74 | 75.2 | |
| Bae et al. (Patch-Mix CL)Training protocol=Patch-Mix CL2026.04 | 70.78 | 70 | 70.39 | |
| Kim et al. (Audio-CLAP)Training protocol=Audio-CLAP2026.04 | 70.58 | 80 | 75.29 | |
| Kim et al. (SG-SCL)Training protocol=SG-SCL2026.04 | 70.22 | 74.5 | 72.36 | |
| Bae et al. (Fine-tuning)Training protocol=Fine-tuning2026.04 | 69.47 | 67.5 | 68.49 |