Complex Gaussian Mixture Model for fingerprint minutiae

Fingerprint representation is important in fingerprint recognition systems and has great impact on its performance. In this paper, we first introduce complex continuous density functions named Complex Gaussian Mixture Model (CGMM) to represent the fingerprint minutiae. In this model, a Gaussian mixture model is constructed according to the positions of fingerprint minutiae, and the direction of minutiae is joined to each Gaussian as a complex composition. Then the similarity of two aligned CGMM is defined by the correlation intensity between them and can be simplified to an equivalency summation through strict mathematical derivation. Finally the CGMM model and the similarity measurement method are applied in fingerprint matching and retrieval. Experimental results demonstrate that the proposed CGMM model and the similarity measurement method are effective and efficient.