Research Article
Fingerprint Classification Combining Curvelet Transform and Gray-Level Cooccurrence Matrix
Table 4
Comparison of accuracy on NIST-4 for the five-class classification.
| | Features | Accuracy
A | Accuracy
T | Accuracy
L | Accuracy
R | Accuracy
W |
| Wavelet
(Symmlet 4) | 92% | 79.7% | 94.2% | 93.3% | 93.7% | Wavelet
(Symmlet 5) | 92% | 82.7% | 94.5% | 93.5% | 93.1% | Wavelet
(Symmlet 6) | 92% | 83.7% | 94.2% | 93.0% | 93.7% | Wavelet
(Symmlet 8) | 92% | 82.7% | 94.2% | 93.0% | 93.7% | Wavelet
(Symmlet 9) | 92% | 81.2% | 94.2% | 93.0% | 93.7% | | GLCM | 78.4% | 77.7% | 79.4% | 72.8% | 82.7% | | Curvelet | 93.6% | 84.7% | 95.0% | 94.0% | 94.5% | | Proposed algorithm | 96.0% | 88.1% | 96.6% | 96.3% | 96.2% |
|
|
