Template-Driven Knowledge Distillation for Compact and Accurate Periocular Biometrics Deep-Learning Models

Damer, Naser; Raja, Kiran B.; Kirchbuchner, Florian; Kuijper, Arjan

doi:10.3390/s22051921

Cited by 12 publications

(3 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In another study, Boutros et al [47] proposed a novel template-driven KD approach that optimized the distillation process so that the student model learned to produce templates similar to those produced by the teacher model. This was obtained by introducing an additional loss function over the original KD loss operating on the feature extraction layer.…”

Section: Prior Workmentioning

confidence: 99%

Benchmarking Neural Network Compression Techniques for Ocular-Based User Authentication on Smartphones

Almadan

Rattani

2023

IEEE Access

View full text Add to dashboard Cite

With the unprecedented mobile technology revolution, mobile devices have transcended from being the primary means of communication to an all-in-one platform. Consequently, an increasing number of individuals are accessing online services for e-commerce and banking via smartphones instead of traditional desktop computers. However, smartphones can be easily misplaced, lost, or stolen more often than other computing devices, thereby demanding effective user authentication mechanisms for device unlocking and secured transactions. Ocular biometrics has obtained significant attention from academia and industry because of its accuracy, security, and ease of use in mobile devices. Several studies have demonstrated the efficacy of deep learning models for ocular-based user authentication on smartphones. However, these high-performing models require enormous space and computational complexity due to the millions of parameters and computations involved. These requirements make their deployment on resourceconstrained smartphones challenging. To this end, a handful of studies have been proposed for compact-size ocular-based deep-learning models to facilitate on-device deployment. In this paper, we conduct a thorough analysis of the existing neural network compression techniques applied as a standalone and in combination for ocular-based user authentication. Extensive experimental validation is performed on the two latest largescale ocular biometric datasets collected using smartphones, namely, UFPR and VISOB 2.0 datasets. This study benchmarks the results of advanced compression techniques for further research and development in lightweight models for ocular-based user authentication on smartphones.

show abstract

Section: Prior Workmentioning

confidence: 99%

Benchmarking Neural Network Compression Techniques for Ocular-Based User Authentication on Smartphones

Almadan

Rattani

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The biometric literature refers to the recognition of this area, when the iris is not exclusively targeted, as periocular recognition [106]. Periocular recognition can include the periocular region of one eye for some applications [107,108]. However, in the masked face recognition scenario, both right and left periocular regions are typically considered.…”

Section: Enhancing Masked Face Recognitionmentioning

confidence: 99%

A Survey on Computer Vision based Human Analysis in the COVID-19 Era

Eyiokur¹,

Kantarcı²,

Erakin³

et al. 2022

Preprint

View full text Add to dashboard Cite

The emergence of COVID-19 has had a global and profound impact, not only on society as a whole, but also on the lives of individuals. Various prevention measures were introduced around the world to limit the transmission of the disease, including face masks, mandates for social distancing and regular disinfection in public spaces, and the use of screening applications. These developments also triggered the need for novel and improved computer vision techniques capable of (i) providing support to the prevention measures through an automated analysis of visual data, on the one hand, and (ii) facilitating normal operation of existing vision-based services, such as biometric authentication schemes, on the other. Especially important here, are computer vision techniques that focus on the analysis of people and faces in visual data and have been affected the most by the partial occlusions introduced by the mandates for facial masks. Such computer vision based human analysis techniques include face and face-mask detection approaches, face recognition techniques, crowd counting solutions, age and expression estimation procedures, models for detecting facehand interactions and many others, and have seen considerable attention over recent years. The goal of this survey is to provide an introduction to the problems induced by COVID-19 into such research and to present a comprehensive review of the work done in the computer vision based human analysis field.Particular attention is paid to the impact of facial masks on the performance of various methods and recent solutions to mitigate this problem. Additionally, a detailed review of existing datasets useful for the development and evaluation of methods for COVID-19 related applications is also provided. Finally, to help advance the field further, a discussion on the main open challenges and future research direction is given at the end of the survey. This work is intended to have a broad appeal and be useful not only for computer vision researchers but also the general public.

show abstract

“…KD transfers the acquired knowledge learned by a larger network to a smaller one [37], [38]. KD has shown great success in improving the verification performance of compact FR models [10], [11], [39]. Furthermore, the combination of KD with other model compression techniques such as compact model design [11], or NAS [10], [17] demonstrated very promising accuracies in FR.…”

Section: Introductionmentioning

confidence: 99%

QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization

Damer¹,

Kuijper²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Deep learning-based face recognition models follow the common trend in deep neural networks by utilizing full-precision floating-point networks with high computational costs. Deploying such networks in use-cases constrained by computational requirements is often infeasible due to the large memory required by the full-precision model. Previous compact face recognition approaches proposed to design special compact architectures and train them from scratch using real training data, which may not be available in a real-world scenario due to privacy concerns. We present in this work the QuantFace solution based on low-bit precision format model quantization. QuantFace reduces the required computational cost of the existing face recognition models without the need for designing a particular architecture or accessing real training data. QuantFace introduces privacy-friendly synthetic face data to the quantization process to mitigate potential privacy concerns and issues related to the accessibility to real training data. Through extensive evaluation experiments on seven benchmarks and four network architectures, we demonstrate that QuantFace can successfully reduce the model size up to 5x while maintaining, to a large degree, the verification performance of the full-precision model without accessing real training datasets. All training codes are publicly available 1 .

show abstract

Template-Driven Knowledge Distillation for Compact and Accurate Periocular Biometrics Deep-Learning Models

Cited by 12 publications

References 47 publications

Benchmarking Neural Network Compression Techniques for Ocular-Based User Authentication on Smartphones

Benchmarking Neural Network Compression Techniques for Ocular-Based User Authentication on Smartphones

A Survey on Computer Vision based Human Analysis in the COVID-19 Era

QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization

Contact Info

Product

Resources

About