VarGFaceNet: An Efficient Variable Group Convolutional Neural Network for Lightweight Face Recognition

Yan, Mengjia; Zhao, Mengao; Xu, Zining; Zhang, Qian; Wang, Guoli; Su, Zhizhong

doi:10.1109/iccvw.2019.00323

Cited by 86 publications

(72 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lightweight neural networks tailored specifically for the face recognition task have been proposed, such as Mo-bileFaceNet [13], ShuffleFaceNet [66], and VarGFaceNet [107].…”

Section: ) Lightweight Convolutional Neural Network For Face Recognitionmentioning

confidence: 99%

A Study on the Performance of Unconstrained Very Low Resolution Face Recognition: Analyzing Current Trends and New Research Directions

et al. 2021

View full text Add to dashboard Cite

In the past decade, research in the face recognition area has advanced tremendously, particularly in uncontrolled scenarios (face recognition in the wild). This advancement has been achieved partly due to the massive popularity and effectiveness of deep convolutional neural networks and the availability of larger unconstrained datasets. However, several face recognition challenges remain in the context of very low resolution homogeneous (same domain) and heterogeneous (different domain) face recognition. In this survey, we study the seminal and novel methods to tackle the very low resolution face recognition problem and provide an in-depth analysis of their design, effectiveness, and efficiency for a real-time surveillance application. Furthermore, we analyze the advantage of employing deep learning convolutional neural networks, while presenting future research directions for effective deep learning network design in this context.

show abstract

“…Lightweight neural networks tailored specifically for the face recognition task have been proposed, such as Mo-bileFaceNet [13], ShuffleFaceNet [66], and VarGFaceNet [107].…”

Section: ) Lightweight Convolutional Neural Network For Face Recognitionmentioning

confidence: 99%

A Study on the Performance of Unconstrained Very Low Resolution Face Recognition: Analyzing Current Trends and New Research Directions

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Depending on depth of network and spatial resolution of features where such distillation is applied, it makes student to mimic teacher at different levels of abstraction. However, over-reguralization of hidden layers can lead to poor quality, so usually hints are only used for embedding (pre-classification) layer [16,21]. In order to successfully guide student even at initial layers, modification of hints idea was proposed in [13].…”

Section: Individual Knowledge Distillationmentioning

confidence: 99%

“…Some recent works [21,22] follow the idea of hint connections and impose constraints on the discrepancy between teacher's and student's embeddings. But in order to better fit angular nature of conventional losses used to train face recognition networks [28,29,31], authors put penalty on cosine similarity, instead of Euclidean distance.…”

Section: Knowledge Distillation For Face Recognitionmentioning

confidence: 99%

Pairwise Ranking Distillation for Deep Face Recognition

Nikitin¹,

Konushin²,

Konushin

2020

Proceedings of the 30th International Conference on Computer Graphics and Machine Vision (GraphiCon 2020). Part 2

View full text Add to dashboard Cite

This work addresses the problem of knowledge distillation for deep face recognition task. Knowledge distillation technique is known to be an effective way of model compression, which implies transferring of the knowledge from high-capacity teacher to a lightweight student. The knowledge and the way how it is distilled can be defined in different ways depending on the problem where the technique is applied. Considering the fact that face recognition is a typical metric learning task, we propose to perform knowledge distillation on a score-level. Specifically, for any pair of matching scores computed by teacher, our method forces student to have the same order for the corresponding matching scores. We evaluate proposed pairwise ranking distillation (PWR) approach using several face recognition benchmarks for both face verification and face identification scenarios. Experimental results show that PWR not only can improve over the baseline method by a large margin, but also outperforms other score-level distillation approaches.

show abstract

“…Convolutional Neural Networks combines three basic architectures, namely local receptive fields, shared weight in the form of filters, and spatial subsampling in the form of pooling. Convolution or what is commonly known as convolution is a matrix that functions to perform filters [5]. In the filtering process, there are two matrices, namely the input value matrix and the kernel matrix.…”

Section: Introductionmentioning

confidence: 99%

Home Security System with Face Recognition based on Convolutional Neural Network

Irjanto¹,

Surantha²

2020

IJACSA

View full text Add to dashboard Cite

Security of house doors is very important and becomes the basis for the simplest and easiest security and sufficient to provide a sense of security to homeowners and along with technological developments, especially in the IoT field, which makes technological developments in locking house doors have developed a lot like locking house doors with faces and others. The development of facial recognition systems has also developed and has been implemented for home door locking systems and is an option that is quite simple and easy to use and is quite accurate in recognizing the face of homeowners. The development of the CNN method in facial recognition has become one of the face recognition systems that are easy to implement and have good accuracy in recognizing faces and has been used in object recognition systems and others. In this study, using the CNN Alexnet facial recognition system which is implemented in a door locking system, data collection is done by collecting 1048 facial data on the face of the homeowner using a system which is then used to train machine learning where the results are quite accurate where the accuracy is the result is 97.5% which is quite good compared to some other studies. The conclusion is the CNN Alexnet method can perform facial recognition which is quite accurate which can be implemented on the IoT device, namely, the Raspberry Pi.

show abstract

VarGFaceNet: An Efficient Variable Group Convolutional Neural Network for Lightweight Face Recognition

Cited by 86 publications

References 29 publications

A Study on the Performance of Unconstrained Very Low Resolution Face Recognition: Analyzing Current Trends and New Research Directions

A Study on the Performance of Unconstrained Very Low Resolution Face Recognition: Analyzing Current Trends and New Research Directions

Pairwise Ranking Distillation for Deep Face Recognition

Home Security System with Face Recognition based on Convolutional Neural Network

Contact Info

Product

Resources

About