We present a Deep Convolutional Neural Network (DCNN) architecture for thetask of continuous authentication on mobile devices. To deal with the limitedresources of these devices, we reduce the complexity of the networks bylearning intermediate features such as gender and hair color instead ofidentities. We present a multi-task, part-based DCNN architecture for attributedetection that performs better than the state-of-the-art methods in terms ofaccuracy. As a byproduct of the proposed architecture, we are able to explorethe embedding space of the attributes extracted from different facial parts,such as mouth and eyes, to discover new attributes. Furthermore, throughextensive experimentation, we show that the attribute features extracted by ourmethod outperform the previously presented attribute-based method and abaseline LBP method for the task of active authentication. Lastly, wedemonstrate the effectiveness of the proposed architecture in terms of speedand power consumption by deploying it on an actual mobile device.
展开▼
