Abstract
Biometrics has been gaining attraction due to the ever-growing demand of this field of research on access control, public security, forensics and e-banking. However, there are still many challenging problems in improving the accuracy, robustness, efficiency, and user-friendliness of these biometric systems. In this manuscript a new adaptive multi-modal biometric framework based on super Self Organizing Maps (super SOM) for the recognition of individuals using palm print, ear and face is proposed. It is showed that the proposed framework helps to improve the performance and robustness of recognition when compared to standard methods in literature namely Sequential Float Feature Selection and Principal Component Analysis. The major focus of this approach is to keep the framework adaptive and robust, thereby, capable of being used in a wide variety of environments. Moreover some new directions on which super SOM shall be effectively used in biometrics community is also discussed. Towards the end, an arity dimensionality concept (inspired from biology) which further enhances the efficiency of this framework is also used. All the findings are showed with experimental results.