A Combined Feature for Fingerprint Recognition

Partial fingerprint recognition is an important challenge especially when the partial image does not include singular points such as core and delta. In this paper, we propose a new localized feature named combined feature for partial fingerprint recognition. The combined feature combines the information of every two minutiae of the image and the ridges structure between them. This feature is defined based on the minutiae including ridge endings and bifurcations. The combined feature is invariant with respect to the global transformations such as rotation and transformation. The recognition is performed in three steps: minutiae extraction, combined features extraction and matching. Experimental results on FVC2004 show efficiency and accuracy of the proposed method.