Research Of Fingerprint Identification Based On Alignment-free Fingerprint Feature

With the wide application of fingerprint identification system, the precision and speed of automatic fingerprint identification system demands increase. Traditional automatic fingerprint recognition system consists of five modules, including:image acquisition, image preprocessing, feature extraction and matching. But traditional automatic fingerprint identification system has some difficulties in improving system performance. Therefore, we think, we can change the model, optimization of automatic fingerprint identification system, to improve the performance.In this paper we propose a novel alignment-free fingerprint feature, and apply it to automatic fingerprint identification system, in order to ensure the system recognition accuracy and improve the system speed, so as to improve the whole system performance.Considering the changing of fingerprint image, it is usually causes by acquisition of finger position, collecting finger pressure and spot light. This produced displacement, rotation and scale change of fingerprint images. We try to extract feature of fingerprint in frequency domain, and then use the characteristics of Fourier-Mellin transform to eliminate these linear change, obtain the alignment-free feature.Experiments show that, the proposed method can not only save time but also save the storage. Using this feature in automatic fingerprint identification system, can effectively improve the system performance.