Fingerprint Identification using Radon Transform

In this paper, a new method for fingerprint identification based on Radon transform is proposed as an alternative to minutiae-based approaches. Since existing methods based on these local characteristics "minutiae" have attained a level of success rate saturation, this paper aims to open a new gateway by proposing a different approach for analyzing fingerprints. This approach offers a fine analysis from global characteristics via the generated Radon profiles. The basic idea is to exploit the directional properties of Radon transform in order to extract new global features from fingerprint images, which could be considered as a set of oriented patterns. A global structural similarity based on these directional features is then derived. This new approach does not necessitate any pre-processing and treatments as in classical fingerprint identification methods. To evaluate the performance of the proposed method, some fingerprints images are used. The obtained results confirm the efficiency of the method in terms of flexibility, simplicity and quality of identification.