Abstract:
At present, fingerprint verification system has gained increasingly broader applications. The verification can be processed on smart cards which contribute to better security for the system. By transmitting data to be verified onto smartcard, fingerprint template stored inside the smartcard can then be protected from being leaked outside to avoid detection or counterfeit risks. This study has been designed to develop Fingerprint Matching Algorithm for processing directly on 8-bit smartcard having Random Access Memorys capacity less than the entire fingerprint data. This approach seeks to classify fingerprint data already undergone Feature Extraction by dividing them into 3 equal portions for comparison with the fingerprint template stored inside the smartcard, by comparing them with the highest and lowest threshold values. These values are applied to determine whether fingerprint data belong to the same or otherwise different person, and whether additional fingerprint data are required for further verification. The upper one helps lessen verification time should incoming fingerprint images belong to the same person, whereas the lower one helps lessen such time should they belong to different persons having different type and position of minutiae between incoming images and fingerprint template