An Associated Extraction Method of Palmprint Principal Lines

The research of palm print features has attracted a lot of attention, and the principal lines which is one of the stable and important features in palm print images can provide effective information for application of palm print technology. Aimed to accurate and natural extraction, in this paper, an associated extraction method of palm print principal lines is presented based on the own properties of palm print images. The method consists of three steps: coarse segmentation, principal lines position and accurate extraction. Firstly, the coarse segmentation was obtained by Nib lack´s method. Secondly, SUSAN operator was applied to limit the range of principal lines as position result. Finally, the accurate extraction was achieved by the intersection of those two outputs mentioned above. Besides, we also adopted alternating sequential filter to solve thining problem. The experimental results with different shapes of palm print showed that this method is effective for capturing detailed and accurate structural information.