Percussion stroke and tonic classification on Mridangam
97.8Stroke AccuracyHybrid BYOL-S/CvT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Hybrid BYOL-S/CvTalpha:beta ratio=2:12022.06 | 97.8 | 96.6 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=4:12022.06 | 97.8 | 96.4 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=1:12022.06 | 97.7 | 94.9 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=3:22022.06 | 97.7 | 96.7 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=1:22022.06 | 97.6 | 96.7 | |
| BYOL-Sencoder=CvT2022.06 | 97.5 | 94 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=2:32022.06 | 97.5 | 96.6 | |
| BYOL-S++Pre-training dataset=BYOL-S++2022.06 | 97.4 | 91.7 | |
| BYOL-SPre-training dataset=BYOL-S2022.06 | 97.3 | 92.9 | |
| BYOL-Swindow size (s)=0.52022.06 | 97.3 | 93 | |
| BYOL-Swindow size (s)=0.952022.06 | 97.3 | 92.5 | |
| Hybrid BYOL-S/CvTalpha:beta ratio=1:42022.06 | 97.2 | 96 | |
| BYOL-Swindow size (s)=22022.06 | 97.1 | 91.9 | |
| BYOL-APre-training dataset=BYOL-A2022.06 | 97 | 90 | |
| BYOL-Swindow size (s)=1.4252022.06 | 97 | 92.5 | |
| BYOL-S/CvT + OS2022.06 | 96.8 | 93.1 | |
| BYOL-Sencoder=CLSTM2022.06 | 96.7 | 89.7 | |
| BYOL-Sencoder=Resnetish-342022.06 | 96.1 | 88.7 | |
| Hybrid BYOL-S/Default2022.06 | 96.1 | 88.7 | |
| openSMILE (OS)mode=OS only2022.06 | 95.6 | 88 | |
| wav2vec2baseline_type=HEAR baseline2022.06 | 94.7 | 82.8 | |
| CREPEbaseline_type=HEAR baseline2022.06 | 88.7 | 82.3 |