Twins’ biometric fusion and introducing a new dataset

Nowadays using authentication systems for twins' recognition is very important and practical. Uni-biometric systems face some problems like unavailability, noise, the possibility of misleading and lack of uniqueness. So it's better to use a multi-biometric system. Score level fusion has high performance and low time complexity among the existing levels of biometric fusion and it has been always the most interesting level for researchers. Since a complete physiological twins' datasets have not been accessible, we have collected a multi-biometric dataset consist of fingerprint, face and iris biometrics of twins.