A survey on soft Biometrics and their application in person recognition at a distance

Abstract

Soft Biometrics tried to use demographic informations as gender, ethnicity and soft attributes like eye color, height, color, weight and other visible marks like scars, and tattoos as ancillary information to improve the performance of biometric systems. The soft biometric information is extracted from a human body is ancillary information is easy to identify distinguished at a distance but the recognition task is not fully identified. This paper presents the review on some soft biometrics traits informations and how to detect a person at a distance. This soft biometric information is analyzed based on the scenarios of varying distance between camera and subject. The subjects are close, medium and far. The soft biometric label or trait is able to improve the performance of face recognition based on the sparse representation on real and ideal scenarios by adaptive fusion rules.