Biometric recognition of individuals is based on their physiological features such as finger print, face, iris, palm print, DNA and behavior features such as, voice, gait and signature etc The biometric system is categorized into unimodal if single biometric trait is used and multimodal biometric systems if multiple biometric traits are used. Some of the limitation of unimodal biometric system in terms of accuracy, distinctiveness, universality, acceptability etc can be overcome in multi biometric system. This paper proposed new dynamic fusion algorithm to select subset of modalities based on individual characteristics depending upon various conditions and availability of information. Proposed system provide selective approach, this means that fusion of multimodal biometric will be perform only if single biometric trait does not provide good performance for recognition.
Volume 11 | 11-Special Issue