Facial recognition with 2D color barcodes
Marco Querini, Giuseppe Francesco Italiano · Cineca Institutional Research Information System (Tor Vergata University) · 2013
We introduce a novel use of 2D barcodes for storing facial biometrics. To accomplish this, we design 2D color barcodes with larger storage capabilities than traditional 2D barcodes. Biometric data are secured through digital signature so as to be protected from malicious tampering. In order to improve the quality of facial recognition, we combine eigenfaces, shape landmarks identified through statistical models and local features such as SURF descriptors. The use of color barcodes is crucial in this setting, as traditional barcodes cannot store all data required for discriminating faces (e.g., a suitable number of SURF descriptors). We report the results of an experimental evaluation of our system on real-world datasets (i.e., a face database). In spite of the reduced storage available on 2D barcodes, we do not observe any degradation in precision and accuracy.