Soft-biometric attributes from selfie images

Abstract

The aim of this chapter is to discuss the soft-biometric attributes that can be extracted from selfie images acquired from mobile devices. Existing literature suggests that various features in demographics, such as gender and age, in physical, such as periocular and eyebrow, and in material, such as eyeglasses and clothing, have been extracted from selfie images for continuous user authentication and performance enhancement of primary biometric traits. Due to the limited hardware resources, low resolution of front-facing cameras, and the usage of the device in different environmental conditions, factors such as robustness to low-quality data, consent-free acquisition, lower computational complexity, and privacy, favor soft-biometric prediction in mobile devices.