Title:Explainable Biometrics in the Age of Deep Learning

Abstract:Systems capable of analyzing and quantifying human physical or behavioral traits, known as biometrics systems, are growing in use and application variability. Since its evolution from handcrafted features and traditional machine learning to deep learning and automatic feature extraction, the performance of biometric systems increased to outstanding values. Nonetheless, the cost of this fast progression is still not understood. Due to its opacity, deep neural networks are difficult to understand and analyze, hence, hidden capacities or decisions motivated by the wrong motives are a potential risk. Researchers have started to pivot their focus towards the understanding of deep neural networks and the explanation of their predictions. In this paper, we provide a review of the current state of explainable biometrics based on the study of 47 papers and discuss comprehensively the direction in which this field should be developed.