Periocular biometric covering the immediate vicinity of human eye is a synergistic alternative to face particularly when the face is masked or occluded. Most present work for periocular recognition in the wild are mainly convolutional neural networks learned based on cross-entropy loss. However, periocular images only capture the least salient face features, and thus suffering from severe intra-class compactness and inter-class dispersion issues for discriminative deep feature learning. Recently, label smoothing regularization (LSR) is discerned capable of diminishing the intra-class variation by minimizing the Kullback-Liebler divergence of a uniform distribution and a network prediction distribution. In this paper, we extend LSR to that of Generalized LSR (GLSR) by learning a pre-task network prediction, in place of the predefined uniform distribution. Extensive experiments on four periocular in the wild datasets disclose that the GSLR-trained networks prevail over the LSR-based counterpart and other most recent the state of the arts. This is supported by our empirical analyses that the embedding periocular features rendered by GLSR results in better class-wise cluster separation than the conventional LSR.

Open set recognition (OSR) models need not only discriminate between known classes but also detect unknown class samples unavailable during training. One promising approach is to learn discriminative representations over known classes with strong intra-class similarity and inter-class discrepancy. Then, the powerful class discrimination learned from the known classes can be extended to known and unknown classes. Without appropriate regularization, however, the model may learn representations trivially, collapsing unknown class representations to the known class ones. To resolve this problem, we propose Divergent Angular Representation (DivAR) based on two approaches. Firstly, DivAR maximizes its representational discrimination between known classes via a highly discriminative loss. Secondly, to ensure separation between known and unknown classes in the representation space, DivAR boosts the directional variation of representations over global samples. In addition, self-supervision is leveraged to improve the representation’s robustness and extend DivAR to one-class classification. Moreover, unlike other OSR methods that require an extra machinery for inference, DivAR learns and infers in a single module. Extensive experiments on generic image datasets demonstrate the plausibility and effectiveness of DivAR for both OSR and One-Class Classification (OCC) problems.

Over the years, a number of biometric template protection schemes, often based on the notion of “cancelable biometrics”, have been proposed. An ideal cancelable biometric algorithm complies with four criteria: irreversibility, revocability, unlinkability, and performance preservation. Cancelable biometrics employs an irreversible but distance preserving transformation to convert the original biometric templates into protected templates. Matching in the transform domain can be accomplished due to the property of distance preservation. However, distance preservation also entails security issues, a point often overlooked in existing research. In this paper, we have conducted a comprehensive security analysis of distance preservation in cancelable biometrics for the first time. The analysis is based on a pre-image attack, which is launched to break the security of cancelable biometrics under Kerckhoffs’s assumption. Furthermore, we propose a general security analysis framework under the single and cross-transformation attacks, which also employs an information leakage estimation strategy based on mutual information as a complement. The experimental results performed on real face, iris, and fingerprint data demonstrate that the risks originating from the matching scores computed from the distance/similarity of two cancelable templates greatly compromise the security of cancelable biometric schemes, including the classic Biohashing, Index-of-max hashing, Non-linear multi-dimensional spectral hashing, Indexing-First-One hashing, Bloom Filter and Two-factor Protected Minutia Cylinder-Code. The security versus accuracy trade-off is discussed and recommendations for designing a biometric system secure against pre-image attacks are also proposed. The source code is available at github.com/biometricsecurity/CBrisks.