Fusing Dorsal Hand Vein and ECG for personal identification

In this paper, a new approach in human identification is investigated. For this purpose, we fused Dorsal Hand Vein and ECG biometrics to achieve a multimodal biometric system. In the proposed system for fusing biometrics, we used Mel Frequency Coefficient Cepstrum (MFCC) approach in order to extract features of ECG biometric and Line Segment Hausdorff Distance method (LsHD) for matching the vein structures. The Euclidean distance of extracted features from ECG biometric, were measured. At the end, a KNN classifier is applied to identify subject. In order to evaluate the system testing performance, we used two different databases: a dataset of 294 persons in age between 18 and 54 and of different gender, each has 2 images per person, which was acquired at different intervals from their left hand, and the standard PTB database of ECG that contains 549 records from 294 subjects. Moreover, in order to achieve more realistic and reliable results, we gathered Holter ECG recordings acquired from 294 subjects in the same circumstance of the hand image acquisition. The numerical results indicated that the algorithm achieved 94.7% of detection rate.