Table 1.
The MAE of estimated IRRs versus reference data from thermocouple using our proposed algorithm in top five selected locations with lowest errors in the first dataset (Figure 4).
| Subject ID | Error (Location) | ||||
|---|---|---|---|---|---|
| #1 | 0.69 (chest) | 0.72 (neck-R) | 0.25 (arm-L) | 0.43 (finger-L) | 0.66 (wrist-R) |
| #2 | 0.13 (nose) | 0.36 (earlobe-R) | 0.45 (chest) | 3.76 (finger-R) | 0.67 (wrist-bone-R) |
| #3 | 0.44 (nose) | 0.31 (chest) | 2.65 (finger-R) | 1.47 (wrist-L) | 0.22 (wrist-R) |
| #4 | 1.17 (neck-L) | 4.41 (neck-R) | 2.18 (chest) | 3.39 (arm-L) | 4.61 (wrist-L) |
| #5 | 2.31 (earlobe-R) | 5.20 (nose) | 1.65 (neck-L) | 0.52 (neck-R) | 0.49 (chest) |