Table 9 Performance of our models and the notable existing models on the CEDAR and SID dataset.
From: A hybrid machine learning framework for offline signature verification using gray wolf optimization
Ref. | Approach | Performance Metrics |
|---|---|---|
LBP | Accuracy: 98.5% | |
morphological features | Accuracy: 88.41% | |
LBP and LDP with SVM | Accuracy: 86.43% | |
Feature vector extracted from the energy and standard deviation of the curvelet coefficient | Accuracy: 92.17% | |
LBP and ULBP features | Accuracy: 75.53% | |
Agglomerative Hierarchical Clustering | Accuracy: 80% | |
Triangular geometric feature | Accuracy: 66% | |
Uses a combination of local maximum occurrence features and a histogram of orientated Gradient features with the kNN algorithm | Accuracy: 98.4% | |
Uses 22 Gy Level Co-occurrences Matrix (GLCM) and eight geometric features obtained by preprocessing images using SVM | \(\:FAR:\) 1.6; \(\:FRR:\:\)2.1; \(\:AER:\) 1.85 | |
Combined deep and hand-crafted features using autoencoder (a) Fusion of Signet and LBP (b) Fusion of Signet and GLCM | \(\:ERR\): 1.6 (\(\:\pm\:\:0.31\)) | |
VGG16 + RMSProp | Accuracy: 83% | |
Local features | \(\:EER\): 0.2 | |
Global features | \(\:EER\): 0.36 | |
SVM with geometric features | Accuracy: 67.08% | |
SVM with HOG features | Accuracy: 76.67% | |
SignGuard | (a) GWO + CS-LBP features + \(\:\mathcal{H}\mathcal{M}\mathcal{L}\mathcal{F}\); (b) GWO + OC-CSLBP features + \(\:\mathcal{H}\mathcal{M}\mathcal{L}\mathcal{F}\) | (a) 97.46% (b) 98.77% |
