الفهرس Only 14 pages are availabe for public view
 |  | from | 115 |  |  |
| | | from | 115 | | |
Abstract
CONCLUSION AND FUTURE WORK
6.1 Conclusion
• Fingerprint recognition has a good balance of all the desirable properties required for a biometric system. Fingerprints are distinctive, fingerprint details are permanent and fingerprint sensors can easily capture high quality images.
• Wavelet transform can be used for fingerprint feature extraction, this is because it analyzes the image in both space and frequency domains.
• Fourier transform is not a suitable feature extraction technique for fingerprints. This is because it analyzes the image in frequency domain only.
• Feed-forward back propagation networks can be used for fingerprint recognition with good accuracy. They depend on minimizing the error between the output and target feature vectors.
• Correlation technique can be used to solve the problem of capturing the fingerprint with different impressions. The similarity between the different impressions for the same person results in a correlation coefficient value close to one.
• The distance measure classifier can be used to solve the problem of different impressions for the same person. The distance measure value is approximately zero between the different impressions for the same person.
• Minutia-based methods suffer from the problem of false or missed minutia. On the other hand, wavelet transform deals with the rich gray information available in fingerprints, so it is easier and more accurate.
6.2 Future Work
There comes a stage in the development of any biometric recognition system where it becomes increasingly difficult to achieve significantly better performance from a given biometric identifier and the need to explore other sources for improvement becomes a practical necessity. This implies that for the desired performance improvement, we may need to rely on integrating multiple biometrics.