Distillation of an End-to-End Oracle for Face Verification and Recognition Sensors

Abstract

Face recognition functions are today exploited through biometric sensors in many applications, from extended security systems to inclusion devices; deep neural network methods are reaching in this field stunning performances. The main limitation of the deep learning approach is an inconvenient relation between the accuracy of the results and the needed computing power. When a personal device is employed, in particular, many algorithms require a cloud computing approach to achieve the expected performances; other algorithms adopt models that are simple by design. A third viable option consists of model (oracle) distillation. This is the most intriguing among the compression techniques since it permits to devise of the minimal structure that will enforce the same I/O relation as the original model. In this paper, a distillation technique is applied to a complex model, enabling the introduction of fast state-of-the-art recognition capabilities on a low-end hardware face recognition sensor module. Two distilled models are presented in this contribution: the former can be directly used in place of the original oracle, while the latter incarnates better the end-to-end approach, removing the need for a separate alignment procedure. The presented biometric systems are examined on the two problems of face verification and face recognition in an open set by using well-agreed training/testing methodologies and datasets.

A Parameter Refinement Method for Ptychography Based on Deep Learning Concepts

“…To cope with this problem, a spatial transform layer [19] is introduced; the new learnable model is trained by inferring the spatial transformation that, applied to the input feature map, maximises the task metric. This approach greatly improves classification and recognition performances (e.g., in face recognition [46]).…”

“…This approach greatly improves classification and recognition performances (e.g. in face recognition 51 ).…”

Automatic Differentiation for Inverse Problems in X-ray Imaging and Microscopy

“…Calculating derivatives is a need for many numerical techniques that require numerical optimisation. While gradient-less methods [ 16 ] are used to explore solution spaces of a few dimensions (e.g., Bayesian methods to find hyper-parameters), gradients are essential in minimisation problems that involve a huge number of parameters, such as the training of neural networks [ 11 , 14 , 17 ] or a reconstruction of a computational imaging framework. Any use of finite difference methods would result in a very long process.…”

Recent Trends on Applications of Electronics Pervading the Industry, Environment and Society

“…The Special Issue is characterized by 13 original research papers [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 ] that we briefly introduce in the following.…”

“…The eleventh paper [ 11 ] is entitled “Distillation of an End-to-End Oracle for Face Verification and Recognition Sensors” and is authored by F. Guzzi et al, a group of authors from the University of Trieste, the Elettra Sincrotrone Trieste S.C.p.A and the Abdus Salam International Centre for Theoretical Physics.…”