Fig. 1.
Fig. 2.
Fig. 3.
Fig. 4.
Mean accuracy of person identification depending on the number of channels and types of signals used
| ID of base classifier | Components of GRFs used for learning | Accuracy ± SD[%] |
|---|---|---|
| 1 | FL_ML | 89,4314 ± 2.4037 |
| 2 | FL_AP | 91,8562 ± 2.8214 |
| 3 | FL_V | 94,3144 ± 1.5766 |
| 4 | FR_ML | 85,0836 ± 2.7157 |
| 5 | FR_AP | 90,0502 ± 1.9168 |
| 6 | FR_V | 88,0936 ± 3.2901 |
| 7 | FL_ML, FR_ML | 93,1271 ± 1.3641 |
| 8 | FL_AP, FR_AP | 91,3712 ± 2.8435 |
| 9 | FL_V, FR_V | 94,6488 ± 1.3025 |
| 10 | FL_AP, FL_V | 95.4849 ± 2.0797 |
| 11 | FR_AP, FR_V | 92.6756 ± 2.8579 |
| 12 | FL_ML, FL_AP, FL_V | 96,2876 ±1.2359 |
| 13 | FR_ML, FR_AP, FR_V | 94,2642 ±1.1316 |
| 14 | FL_ML, FL_AP, FR_ML, FR_AP | 95.5686 ±1.2014 |
| 15 | FL_AP, FL_V, FR_AP, FR_V | 96.2709 ±1.4558 |
| 16 | FL_ML, FL_AP, FR_AP, FR_V | 96.3378 ±1.2819 |
| 17 | All | 96,5719 ± 1.1403 |
Mean accuracy of person identification depending on the base classifiers used
| ID | ID of base classifiers | Accuracy ±SD [%] |
|---|---|---|
| EC_1 | 1+2+3+4+5+6 | 98.7793 ± 0.4321 |
| EC_2 | 1+2+3+4+5+6+17 | 99.2140 ± 0.2736 |
| EC_3 | 7+8+9 | 97.8930 ± 0.8162 |
| EC_4 | 7+8+9+17 | 98.7625 ± 0.6217 |
| EC_5 | 1+2+3+4+5+6+7+8+9 | 99.1973 ± 0.3326 |
| EC_6 | 1+2+3+4+5+6+7+8+9+17 | 99.4147 ± 0.2398 |
| EC_7 | 7+8+9+10+11 | 98.8127 ± 0.5196 |
| EC_8 | 7+8+9+10+11+17 | 99.1639 ± 0.3053 |
| EC_9 | 12+13+17 | 98.4114 ± 0.6884 |
| EC_10 | 1+2+3+4+5+6+7+8+9+10+11 | 99.3645 ± 0.2708 |
| EC_11 | 1+2+3+4+5+6+7+8+9+10+11+17 | 99.4482 ± 0.2848 |
| EC_12 | 1+2+3+4+5+6+7+8+9+10+11+12+13 | 99.3478 ± 0.2423 |
| EC_13 | 1+2+3+4+5+6+7+8+9+10+11+12+13+17 | 99.4314 ± 0.2518 |
| EC_14 | 14+15+16 | 98.1940 ± 0.4303 |
| EC_15 | 14+15+16+17 | 98.7625 ± 0.3629 |
| EC_16 | 1+2+3+4+5+6+7+8+9+10+11+12+13+14+15+16 | 99.5317 ± 0.2238 |
| EC_17 | All base classifiers | 99.5652 ± 0.2257 |
| EC_18 | 2+3+5+7+8+9+10+11+12+13+14+15+ 16+17 | 99.4816 ± 0.2548 |
| EC_19 | 10+12+14+15+16+17 | 99.0635 ± 0.3797 |
The summary of architecture of convolution neural network
| No of layer | No. of Conv Block | Type of layer | Kernel size | No of kernels | Output size |
|---|---|---|---|---|---|
| 1 | - | Input | - | - | 1643 x channels |
| 2 | 1 | Conv1D | 5 | 64 | 1639 x 64 |
| 4 | Max Pooling | 2 | - | 819 x 64 | |
| 5 | 2 | Conv1D | 3 | 128 | 817x128 |
| 7 | Max Pooling | 2 | - | 408x128 | |
| 8 | 3 | Conv1D | 3 | 256 | 406x256 |
| 10 | Max Pooling | 2 | - | 203x256 | |
| 11 | 4 | Conv1D | 3 | 512 | 201x512 |
| 13 | Max Pooling | 2 | - | 100x512 | |
| 14 | 5 | Conv1D | 3 | 1024 | 98x1024 |
| 15 | Max Pooling | 2 | 49x1024 | ||
| 16 | - | Flatten | - | - | 50 176 |
| 17 | - | Fully-Connected1 | - | 1000 neurons | 1000 |
| 18 | - | Fully-Connected2 | - | 700 neurons | 700 |
| 19 | - | Output | - | 322 neurons | 322 |