Table 5.
Performance comparison with other methods for SCMI (Best results are highlighted in bold).
| Camera Model | Identification Accuracy (%) | |||||
|---|---|---|---|---|---|---|
| Frame-Level | Video-Level | |||||
| MISLnet [43] | PRNU-Net [51] | Our Method | MISLnet [43] | PRNU-Net [51] | Our Method | |
| Huawei Y7 | 70.4 | 71.5 | 98.4 | 75.8 | 86.0 | 99.2 |
| Huawei Y9 | 76.9 | 77.3 | 99.3 | 90.8 | 91.6 | 100.0 |
| Nokia 5.4 | 81.6 | 83.1 | 90.3 | 91.7 | 92.7 | 90.8 |
| Nokia 7.1 | 76.7 | 80.6 | 96.6 | 79.8 | 92.2 | 96.7 |
| Redmi Note 8 | 72.8 | 81.9 | 95.2 | 84.2 | 84.2 | 95.0 |
| Redmi Note 9 Pro | 65.0 | 75.0 | 91.7 | 70.0 | 77.3 | 93.3 |
| Samsung A50 | 68.8 | 75.0 | 96.8 | 70.0 | 77.2 | 96.7 |
| Samsung Note 9 | 72.0 | 78.8 | 92.4 | 87.6 | 95.8 | 93.3 |
| iPhone 8 Plus | 69.6 | 73.9 | 91.2 | 68.3 | 85.5 | 94.2 |
| iPhone XS Max | 69.3 | 79.9 | 84.7 | 70.0 | 74.9 | 85.8 |
| Average Accuracy | 72.5 | 77.7 | 93.6 | 78.8 | 85.74 | 94.5 |