A Review of Face Recognition Methods Based on Lightweight Deep Neural Networks

Face recognition has always been a research hotspot in the field of artificial intelligence. Although the most advanced methods are mainly deep learning methods based on convolutional neural networks, the huge computational overhead is also the main factor restricting its application. While improving the accuracy rate, reducing the computing and storage overhead of deep neural networks is a very important and challenging problem in the field of face recognition. Therefore, it is very important to study the face detection and recognition method based on lightweight network and design highly discriminative feature representation. This paper first introduces the recognition methods of face recognition, including traditional recognition methods and quantified recognition methods, and summarizes the advantages and disadvantages of face recognition. Then the representative data sets and commonly used evaluation indexes in the field of lightweight neural networks are combed. Finally, the development trend of lightweight neural network face recognition technology is forecasted.