Table 8.
Comparison between the recognition method suggested by this study and the methods of other preceding studies with the MICHE database including the data from indoors and outdoors (N.R. means “not reported”).
| Method | Sub-Dataset | EER (%) | d-Prime Value |
|---|---|---|---|
| Abate et al.’s [67] method | Galaxy S4 | 36.7 | 0.65 |
| Galaxy Tab2 | 39.1 | 0.60 | |
| iPhone5 | 39.9 | 0.51 | |
| Barra et al.’s [68] method | Galaxy S4 | 45 (approximate value) | N.R. |
| Galaxy Tab2 | 46 (approximate value) | ||
| iPhone5 | 43 (approximate value) | ||
| Raja et al.’s [28] method * | Galaxy S4 | 38.8 | 6.49 |
| Galaxy Tab2 | 33.9 | 8.63 | |
| iPhone5 | 38.6 | 6.21 | |
| Santos et al.’s [69] method ** | Galaxy S4 | 19.8 | 6.13 |
| Galaxy Tab2 | 16.3 | 6.20 | |
| iPhone5 | 22 | 5.44 | |
| Proposed method | Galaxy S4 | 17.9 | 1.87 |
| Galaxy Tab2 | 16.25 | 2.26 | |
| iPhone5 | 17.45 | 2.00 |
* and **: the accuracies are reported in the study by De Marsico et al. [70].