Figure 1.
Face Recognition Process
Figure 2.
Face Recognition Stages
Figure 3.
Facial Detection and Alignment
Figure 4.
Facial Detection by Different Detectors
Figure 5.
Nodal Points on Two Different People
Table 1.
Cosine metric accuracy
| Detector | FaceNet5l2 | Face Net | VGG-Face | ArcFace | SFace | DeepFace | DeepID |
|---|---|---|---|---|---|---|---|
| retinaface | 98.4 | 96.4 | 95.8 | 96.6 | 92.4 | 67.7 | 64.4 |
| mtcnn | 97.6 | 96.8 | 95.9 | 96 | 90.5 | 66.3 | 63 |
| fastmtcnn | 98.1 | 97.2 | 95.8 | 96.4 | 90 | 67.4 | 63.6 |
| dlib | 97 | 92.6 | 94.5 | 95.1 | 69.8 | 66.5 | 58.7 |
| yolov8 | 97.3 | 95.7 | 95 | 95.5 | 91.9 | 67.5 | 65.9 |
| yunet | 97.9 | 97.4 | 96 | 96.7 | 91 | 66.5 | 63.5 |
| centerface | 97.7 | 96.8 | 95.7 | 96.5 | 89.3 | 67.8 | 63.6 |
| mediapipe | 96.1 | 90.6 | 92.9 | 90.3 | 75.4 | 64.8 | 63 |
| ssd | 88.7 | 87.5 | 87 | 86.2 | 84.5 | 63.8 | 62.6 |
| opencv | 87.6 | 84.9 | 87.2 | 84.6 | 83.6 | 63.7 | 60.1 |
Table 2.
Euclidean metric accuracy
| Detector | FaceNet5l2 | FaceNet | VGG-Face | ArcFace | SFace | DeepFace | DeepID |
|---|---|---|---|---|---|---|---|
| Retinaface | 95.9 | 93.5 | 95.8 | 85.2 | 80.2 | 67 | 65.6 |
| MTCNN | 95.2 | 93.8 | 95.9 | 83.7 | 77.4 | 66.5 | 63.5 |
| FastMTCNN | 96 | 93.4 | 95.8 | 83.5 | 77.7 | 66.7 | 64 |
| Dlib | 96 | 90.8 | 94.5 | 88.6 | 66.3 | 63.4 | 60.4 |
| Yolov8 | 94.4 | 91.9 | 95 | 84.1 | 73.4 | 69 | 66.5 |
| Yunet | 97.3 | 96.1 | 96 | 84.9 | 79.4 | 65.8 | 65.2 |
| CenterFace | 97.6 | 95.8 | 95.7 | 83.6 | 77.4 | 65.5 | 62.8 |
| MediaPipe | 95.1 | 88.6 | 92.9 | 73.2 | 72.5 | 61.8 | 62.2 |
| SSD | 88.9 | 85.6 | 87 | 75.8 | 76.9 | 63.4 | 62.5 |
| OpenCV | 88.2 | 84.2 | 87.3 | 73 | 81.1 | 65.5 | 59.6 |
Table 3.
Cosine Similarity distance
| Sample | FaceNet5l2 | FaceNet | VGG-Face | ArcFace | SFace | DeepFace | DeepID | Ensemble |
|---|---|---|---|---|---|---|---|---|
| 20 | 0.19 | 0.12 | 0.46 | 0.35 | 0.29 | 0.11 | 0.12 | 0.12 |
| 50 | 0.17 | 0.2 | 0.42 | 0.31 | 0.22 | 0.2 | 0.27 | 0.2 |
Table 4.
Measure and Declared accuracy
| Detector | FaceNet512 | FaceNet | VGG-Face | ArcFace | SFace | DeepFace | DeepID | Ensemble |
|---|---|---|---|---|---|---|---|---|
| Measured | 98.4 | 97.4 | 96.7 | 96.7 | 93 | 69 | 66.5 | 98.80 |
| Declared | 99.6 | 99.2 | 98.9 | 99.5 | 99.5 | 97.3 | 97.4 | NA |
Figure 6.
Measured and Declared Accuracy Comparision
Figure 7.
Time Taken local and cloud environment.
Figure 8.
Time Taken local and cloud environment.