BioSoft-a multimodal biometric database incorporating soft traits

Abstract

Biometrics is an automated authentication mechanism that allows the identification or verification of individuals based on unique physiological and behavioral characteristics. In addition to novel biometric recognition frameworks and protocols, standard databases containing sample biometric traits are essential for validating the obtained results. In this paper, we introduce a new multimodal database named BioSoft which consists of biometric data collected from 75 individuals. In comparison to the already existing databases, BioSoft contains a set of 23 soft biometric traits corresponding to each enrolled individual. This property makes our database very useful due to the unavailability of any other manually extracted multimodal database incorporating soft biometric characteristics. Additionally, the primary biometric modalities of face, ear, iris, voice, handwriting and fingerprints (obtained from two different sensors) are present in this database. Thus our database contains both physiological and behavioral characteristics of individuals, thus making it applicable for validating a wide variety of approaches. © 2017 IEEE.