Abstract:
The non-intrusive nature and high accuracy of face recognition algorithms have led to their successful deployment across multiple applications ranging from border access to mobile unlocking
and digital payments. However, their vulnerability against sophisticated as well as cost-effective
presentation attack mediums raises essential questions regarding its reliability. Researchers have
proposed multiple presentation attack detection algorithms; however, they are still far behind
from reality. The major problem with the existing work is the generalizability against multiple
attacks both in the seen and unseen setting. The algorithms which are useful for one kind of
attack (such as print) fail miserably for another type of attack (such as silicone masks). The
contributions of this research are two folds. Firstly, we have proposed a deep learning-based network called MixNet to detect presentation attacks in cross-database and unseen attack settings.
The proposed algorithm utilizes state-of-the-art convolutional neural network architectures and
learns the feature mapping for an individual category of attacks. Experiments are performed using two challenging face presentation attack databases, namely, Silicone Mask Attack Database
(SMAD) and Spoof In the Wild with Multiple Attack Types (SiW-M). Extensive experiments
and comparison with hand-crafted features and deep CNN architectures show the effectiveness
of MixNet. Further, this research’s second contribution is a first-of-its-kind face presentation
attack database with multiple variations in both attack and genuine samples. None of the existing databases incorporate the vast cultural diversity found in countries like India. Hence we
propose a novel database that includes multiple variations such as tilak, turban, bindi, etc.