Distance-based feature extraction for biometric recognition of Millimeter Wave body images

摘要

In this work a complete process of feature extraction for biometric recognition of Millimeter Wave body images is described. The scope of this work is to find a small set of distance-based features that can be used in body geometry authentication obtaining good error rates. This approach constitutes a feature-based alternative to the holistic recognition methods already proposed on the mentioned kind of images. The system is tested on a database comprising 1200 synthetic images at 94 GHz of the body of 50 individuals. The results prove that the use of a small number distance-based features provide good class separation.