Studies Advanced in Mask Face Recognition based on Deep Learning

Muheng He

doi:10.54097/hset.v39i.6728

Authors

Muheng He

DOI:

https://doi.org/10.54097/hset.v39i.6728

Keywords:

Face Recognition; Mask; Deep Learning.

Abstract

In the era of epidemic, wearing a mask when going out has become a more common phenomenon. And how to complete face recognition and guarantee certain accuracy without letting users take off their masks has been a challenge in recently years’ face recognition field. In this paper, the traditional masked face recognition method and the new MTCNN deep learning masked face recognition method are compared and analyzed, and other similar masked experiments are introduced as data references. It is synthetically found that the accuracy of MTCNN masked face recognition using CASIA-FaceV5 as dataset is more than 90%, which is far better than other traditional recognition methods, but its requirement for dataset image quality is high, and the training time is long and costly. We hope that the process analysis and result comparison of different methods in this paper can shed light on the future development of masked face recognition technology.

Downloads

Download data is not yet available.

References

Ahmadlou M, Adeli H. Enhanced probabilistic neural network with local decision circles: A robust classifier[J]. Integrated Computer Aided Engineering, 2010, 17(3):197-210.

Jürgen Schmidhuber. Deep learning in neural networks[J]. Neural Netw, 2015.

Zhang K, Zhang Z, Li Z, et al. Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks[J]. IEEE Signal Processing Letters, 2016, 23(10):1499-1503.

Sandler M, Howard A, Zhu M, et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2018.

Schroff F, Kalenichenko D, Philbin J. FaceNet: A Unified Embedding for Face Recognition and Clustering: IEEE, 10.1109/CVPR.2015.7298682[P]. 2015.

Chai Y, Liu J, Li Y. Facial Target Detection and Keypoints Location Study Using MTCNN Model[J]. Journal of Physics: Conference Series, 2021, 2010 (1):012097.

King D E. Dlib-ml: A Machine Learning Toolkit[J]. Journal of Machine Learning Research, 2009, 10(3):1755-1758.

Tang P, Wang H, Kwong S. G-MS2F: GoogLeNet based multi-stage feature fusion of deep CNN for scene recognition[J]. Neurocomputing, 2016, 225(feb.15):188-197.

Umer S, Dhara B C, Chanda B. Face recognition using fusion of feature learning techniques[J]. Measurement, 2019, 146:43-54.

He E J. Masked correlation filters for partially occluded face recognition[C]//IEEE International Conference on Acoustics. IEEE, 2016.