BIOMETRIC INFORMATION FUSION

Due to some problems associated with the unimodal biometric data, such as small variation over the population, large intra-variability over time, absence of biometric sample in portion of a population etc., the use of multimodal biometrics is a first choice solution. The main objective of a multimodal biometric system is to improve the recognition performance of the system and to make the system robust over the limitations associated with unimodal biometric systems. Over the years, several approaches have been proposed and developed for multimodal biometric authentication system with different biometric traits and with different fusion mechanisms.

Multimodal biometric systems use multiple sources of biometric information, whereas information fusion is essential for analysis, indexing and retrieval of such information. There are numbers of fusion techniques for any particular information. Choosing appropriate fusion techniques for any specific information depends on the necessity of the application and the performance of the fusion techniques proven by previous research. There is a consensus in biometric literature that all various levels of multimodal biometric information fall into two broad categories: before matching and after matching fusion. Fusion before matching category contains sensor level fusion and feature level fusion, while fusion after matching contains match score level fusion, rank level fusion and decision level fusion. A novel fusion mechanism has been established recently in BT Lab is based on fuzzy logic fusion, and hence named a fuzzy biometric fusion. Fuzzy biometric fusion can be employed either in the initial stage, i.e. before matching occurred or in the latter stage, i.e. after matching occurred.

Fusion in this category integrates evidences before matching or comparison of data samples against the user sample occurs. According to Kokar et al., “By combining low level features it is possible to achieve a more abstract or a more precise representation of the world”. Thus, biometric sources at the earlier stage contain much more information than after processing).

However, the extra costs of storing raw data and additional complexity in developing matching methods do not make this approach quite practical.

Fusion after-matching methods consolidate information obtained after individual biometric matching or comparison is done. Most multimodal biometric systems have been using these fusion methods as the information needed for fusion is easily available compared to fusion before matching methods. The matching scores, the ranking list (sorted order) based on matching scores or the individual biometric decision (Yes/No) can be used for fusion in this category.