Research Article
SSL: A Novel Image Hashing Technique Using SIFT Keypoints with Saliency Detection and LBP Feature Extraction against Combinatorial Manipulations
Table 3
of the proposed algorithm and other algorithms under different manipulations when
.
| Manipulation | R&A SCH hashing [10] | SIFT-QZMs hashing [25] | SVD-CSLBP hashing [36] | GF-LVQ hashing [37] | RP-NMF hashing [38] | Proposed hashing |
| Brightness adjustment | 1.0 | 1.0 | 1.0 | 0.954 | 1.0 | 1.0 |
| Contrast adjustment | 1.0 | 1.0 | 0.996 | 0.902 | 0.977 | 1.0 |
| Gamma correction | 0.959 | 1.0 | 0.947 | 0.677 | 0.905 | 1.0 |
| Gaussian low-pass filtering | 0.991 | 1.0 | 1.0 | 0.981 | 1.0 | 0.987 |
| Salt and pepper noise | 0.930 | 0.979 | 0.993 | 0.979 | 0.986 | 0.996 |
| Multiplicative noise | 0.957 | 0.984 | 0.985 | 1.0 | 0.991 | 0.995 |
| JPEG compression | 0.963 | 1.0 | 1.0 | 0.795 | 1.0 | 1.0 |
| Watermark embedding | 1.0 | 1.0 | 0.986 | 0.812 | 0.962 | 1.0 |
| Scaling | 0.872 | 0.962 | 1.0 | 0.834 | 0.956 | 0.953 |
| Rotation | 0.878 | 0.835 | 0.773 | 0.596 | 0.843 | 0.919 |
| Double manipulations | 0.83 | 0.824 | 0.753 | 0.71 | 0.817 | 0.893 |
| Triple manipulations | 0.766 | 0.795 | 0.687 | 0.546 | 0.78 | 0.850 |
| Quadruple manipulations | 0.657 | 0.664 | 0.534 | 0.328 | 0.458 | 0.702 |
|
|