OcularNet: Deep Patch-based Ocular Biometric Recognition

Deep learning played a major role in many recent advancements in mobile ocular biometrics. However, many of the experiments are conducted on models which are large in the number of parameters and are inefficient to deploy on mobile devices. In this paper we propose OcularNet, a convolutions neural network(CNN) model, using patches from the eye images. In OcularNet model, we extract six registered overlapping patches from the ocular and periocular region and train a small convolutions neural network(CNN) for each patch named PatchCNN to extract feature descriptors. As the proposed method is a patch-based technique, one can extract features based on the availability of the region in the eye image. We compare verification performance of the proposed Ocular- Net which has 1.5 M parameters with the popular ResNet-50 model which has 23.4 M parameters. On popular large-scale mobile VISOB dataset, the proposed OcularNet model not only outperformed ResNet-50 with at least 11% GMR at 1
^-4
, FMR in subject independent verification setting but also has 15.6 X less number of parameters. Further, experimental evaluations were performed on UBIRIS-I, UBIRIS-II, and CROSS-EYED datasets to evaluate the performance of OcularNet over ResNet-50.