A Fast Fingerprint Indexing Technique with Least Information using Distance Feature

Objectives: To develop a novel fingerprint indexing system which is robust against rotation, scaling and noise? Method/ Analysis: Indexing in large database is a challenging problem as it reduces the number of comparisons. This paper presents a new way of fingerprint database indexing technique using the Euclidian distance from core to minutiae points. Then a set of tuples is created by selecting twenty nearest minutiae points from the core. It performs consistently from image distortion to rotation invariance. It requires less space as it deals only with numerical value and also considers error tolerance “k” as the fingerprint has elastic property. The experimentation is done on FVC 2002 dataset. The high hit rate achieved at low penetration rate indicates that the proposed distance feature indexing technique is satisfactory. Findings: The proposed system can able to retrieve fingerprint records with high hit rate at low penetration rate. Novelity/Improvement: The proposed system uses the average values of the first three tuples and the last three tuples from the selected set of tuples for indexing.