The large amount of research on multimodal systems raises an important question: can we extract additional information fromation from unimodal systems? In this paper, we propose a rank-based score normalization framework that addresses this problem when multi-sample galleries are available. The main idea is to partition the matching scores into subsets and normalize each subset independently. In addition, we present two versions of our framework that: (i) use gallery-based information (i.e., gallery versus gallery scores), and (ii) update available information in an online fashion. We use the theory of Stochastic Dominance to illustrate that the proposed framework can increase the system's performance. Our approach: (i) does not require tuning of any parameters, (ii) can be used in conjunction with any score normalization technique and any integration rule, and (iii) extends the use of W-score normalization to multi-sample galleries. While our approach is better suited for an Open-set Identification task, we also demonstrate that it can be used for a Verification task. In order to assess the performance of the proposed framework we conduct experiments using the BDCP Face database. Our approach improves the Detection and Identification Rate by 14.87% for Z-score and by 4.82% for W-score.
展开▼
