Deep Learning in Biometrics

The course aims at presenting biometric recognition approaches based on recent deep learning techniques. In traditional biometric systems, the designer needs to develop algorithms to extract a set of discriminative features from data. Deep learning approaches learn sets of discriminative features directly from multidimensional signals. The course presents biometric systems from a technological point of view and provides an excursus of deep learning approaches, including: Convolutional Neural Networks, Autoencoders, and Deep Belief Neural Networks. Outline: 1) Biometric systems; 2) Deep learning; 3) Biometric systems based on deep learning techniques; 4) Design and implementation of biometric systems based on deep learning techniques.