Research on the Management of Building Construction Based on Face Recognition
Feng Wang · 2022
In order to improve the attendance management level of construction personnel, a face recognition algorithm based on additive angular boundary loss was proposed to further improve the accuracy of face recognition. First, the recognition algorithm based on additive angular boundary loss was selected as the face recognition algorithm. Then, ResNet50 neural network was selected for face feature extraction. Experimental results showed that the face recognition algorithm based on additive angular boundary loss can maintain high recognition accuracy, and the recognition accuracy on LFW and YTF datasets reaches 99.8% and 98.3%, respectively. After optimization, the accuracy of ResNet50 neural network is also improved when extracting feature information. The above experimental results verify the rationality of the improvement of the recognition algorithm and feature extraction network, which has certain reference value and significance.