Multi-DS Strategy for Source Camera Identification in Few-Shot Sample Data Sets
Table 8
Confusion matrix of the results of multi-DS (%) (VISION database).
Sa1
Ap1
Hu1
Lg1
Ap2
Ap3
Le1
Sa2
Ap4
Mi1
Sa1
93.5
0.1
2.5
0.9
0.4
1.6
0.8
0.3
—
—
Ap1
—
76.2
1.9
1.2
12.0
4.0
2.6
0.9
—
1.1
Hu1
1.2
2.6
87.7
2.9
0.2
2.1
3.2
—
—
—
Lg1
0.1
0.3
2.6
90.9
1.0
1.3
3.6
—
—
0.1
Ap2
0.1
8.9
0.7
1.2
84.0
3.0
0.7
0.6
0.1
0.7
Ap3
0.6
2.9
2.1
1.8
3.9
86.3
1.7
0.7
—
—
Le1
0.1
0.2
1.5
1.9
0.7
0.2
95.1
—
—
0.4
Sa2
0.9
—
0.8
0.1
—
—
—
98.2
—
—
Ap4
—
0.1
—
—
0.4
0.2
—
—
98.9
0.4
Mi1
0.1
0.5
1.0
1.4
0.2
0.5
0.5
—
0.3
95.6
The bold values (diagonal values) indicate that the prediction labels of the model are equal to the real labels, that is, the accuracy of the correctly classified results.
