An evaluation of video-to-video face verification
IEEE Transactions on Information Forensics and Security
Extracting discriminative color features for face recognition
Pattern Recognition Letters
Discriminant analysis and similarity measure
Pattern Recognition
Hi-index | 0.00 |
The paper presents a novel way of matching score calculation for LDA-based face verification. Different from the classical matching schemes, where the decision regarding the identity of the user currently presented to the face verification system is made based on the similarity (or distance) between the "live" feature vector and the template of the claimed identity, we propose to employ a measure we named the corrected normalized correlation coefficient, which considers both the similarity with the template of the claimed identity as well as the similarity with all other templates stored in the database. The effectiveness of the proposed measure was assessed on the publicly available XM2VTS database where encouraging results were achieved.