Infrared face recognition based on Yolo network and its application

Abstract

In the process of indoor face detection, the problem of uneven illumination angle and illumination is always inevitable, which will have a certain impact on face detection and recognition. Therefore, this paper proposes an infrared face recognition algorithm based on Yolo network. Due to the small indoor face data set, the trained model often over fits the samples in the training set, resulting in poor generalization ability and so on. Firstly, this paper uses data enhancement technology to amplify the limited data set. Then, labelme software is used to label the face of the enhanced data set, and the data set is trained through Yolo network to generate the training model corresponding to the data set. Finally, the training model is invoked to test the infrared face images. The experimental results show that the algorithm can recognize the face quickly and accurately.