Periocular-based soft biometric classification

Soft biometric traits refer to characteristics that provide some information about an individual but do not possess the distinctiveness and permanence necessary to sufficiently differentiate between any two individuals. Examples of soft biometric traits include gender, ethnicity, age, weight, and height. As early as 1997, researchers suggested that soft biometrics could be used to improve biometric recognition performance [1]. Researchers later demonstrated the use of gender, ethnicity, and height to improve the performance of a fingerprint recognition system [2]. This chapter discusses the use of local appearance features extracted from the periocular region for gender and ethnicity classification.