Mask face recognition method, system and equipment based on graph convolution fusion network

Abstract

The invention discloses a mask face recognition method, system and equipment based on a graph convolution fusion network. First, the improved ResNet-50 network is used to perform feature representation on the input mask face image, and at the same time, the face key point detection technology is used to obtain Input the eyebrow RoI information of the face; then input the face feature map and eyebrow RoI information into the eyebrow area pooling module to obtain the local features and global features of the eyebrows, and then obtain the final face discrimination features through the graph convolution fusion network; finally , using the ArcFace loss function to optimize the parameters of the CNN skeleton and the graph convolutional fusion network, resulting in a more discriminative mask face recognition feature. The method of the present invention effectively overcomes the problem of low precision of the existing face recognition method under the condition of being covered by a mask.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and relates to a mask-wearing face recognition method, system and equipment, in particular to a mask face recognition method, system and equipment based on a graph convolution fusion network. technical background [0002] Since the beginning of the epidemic, wearing a mask has been a must for residents of almost all countries and regions to travel. However, severe facial occlusion will pose a serious challenge to face recognition technology. Affected by this, the accuracy of face recognition in scenarios such as mobile phone unlocking, security inspection tickets, gate access, and security monitoring will decline to varying degrees. How to effectively model facial occlusion and complex noise and restore real facial information is an important and challenging problem in robust face recognition. [0003] A study by the National Institute of Standards and Technology (NIST) found that wearin...

AI Technical Summary

Problems solved by technology

However, limited by the way masks are worn and the size of the face database, the existing mask face recognition technology is not reliable, and more effective methods need to be explored for face recognition problems such as wearing masks.

Images