An automatic algorithm for human identification using hand X-ray images

In this work we propose a fully automatic algorithm for human identification using hand X-ray images. More specifically, this approach is appropriate to prevent forgeries in high security level systems. The proposed algorithm consists of three steps, namely: (1) segmentation of phalanges, (2) feature extraction using complex Fourier descriptors, (3) identification based on k-nearest neighbor method. To evaluate the proposed protocol, we have constructed a database containing 32 hand X-ray images, acquired using Apollo EZ X-ray machine from 16 non-pathological adult individuals. Preliminary results show that a 100% identification rate is obtained in some specific conditions.