Table 3.
The public results on the China Physiological Signal Challenge (CPSC) and the Georgia 12-lead ECG challenge (G12EC) test sets and the modified 5-fold cross-validation results (mean ± std) of our method on the training sets of CPSC and G12EC.
| Rank in the Challenge 1 | Team Name | CPSC | G12EC | ||
|---|---|---|---|---|---|
| Accuracy | CWAcc | Accuracy | CWAcc | ||
| 1 | ISIBrno-AIMI | 0.422 | 0.725 | 0.250 | 0.617 |
| 2 | NIMA | 0.591 | 0.772 | 0.253 | 0.620 |
| 3 | DSAIL_SNU | 0.258 | 0.640 | 0.179 | 0.594 |
| 4 | Cardiocallenger | 0.405 | 0.699 | 0.268 | 0.619 |
| 5 | USST_Med | 0.252 | 0.574 | 0.179 | 0.597 |
| 6 | CeZIS | 0.650 | 0.910 | 0.286 | 0.673 |
| 7 | SMS + 1 | 0.370 | 0.638 | 0.175 | 0.554 |
| 8 | DataLA_NUS | 0.455 | 0.650 | 0.307 | 0.630 |
| 9 | Dr_Cubic | 0.585 | 0.845 | 0.280 | 0.637 |
| 10 | Ami_kagoshima | 0.499 | 0.763 | 0.350 | 0.600 |
| - | Ours | 0.633 ± 0.013 | 0.723 ± 0.010 | 0.194 ± 0.004 | 0.580 ± 0.006 |
1 The rankings are based on the cost-weighted accuracy (CWAcc) score calculated on all of the four hidden test sets. Detailed public results can be found at https://physionetchallenges.org/2021/leaderboard/ (accessed on 1 November 2021).