Fingerprint Identification Using Fractal Geometry

Abstract

Fingerprints have been used for personal identification for long time. The fingerprints of any individual are unique and do not change throughout his/her life. The proposal research is using Fractal Geometry as a tool for fingerprint identification. The system has two main phases: Training phase and Test phase. For each phase, the system is applied to two main stages on each fingerprint: preprocessing and features extraction. The preprocessing sub stages consist of some image processing techniques as: Enhancement, Binarization, and Thinning. The extracted attributes are the ridges density, minutiae concentration and the directional intersection count with ridges. From these ridge attributes, the fractal dimension is computed to produce fractal arrays. The test results showed that fractal geometry is a powerful tool in fingerprint recognition, where the attained identification rate is (95%) for training set, while this rate is (92%) for the test set, where the database contains samples for 51 person, 8 samples for each person.