Face Recognition Based on MTCNN and Integrated Application of FaceNet and LBP Method
Zaiye Yang, Wei Ge, Zheng Zhang
Abstract
Technology of face recognition has developed rapidly in the past three decades. Various face recognition methods have been proposed by a lot of research. The primary objective of the development of face recognition is improving the accuracy. Multi-task Cascaded Convolutional Networks (MTCNN) is an effective method to detect faces, which identifies the position of the face in the picture and marks five landmarks through deep Convolutional Neural Network (CNN). FaceNet is a technology of face recognition, which is also based on CNN technology, exhibit high accuracy. Local Binary Pattern (LBP) is a traditional technology of face recognition. Despite a lower accuracy than FaceNet, it has many advantages such as grayscale invariance and illumination insensitivity. In this research, we propose an enhanced model of face recognition which is based on MTCNN and integrated application of FaceNet and LBP method. The work that described in this article using LBP parallel FaceNet to improve the illumination robustness of the model only consists of MTCNN and FaceNet. Experiments show that the enhanced model is very effective in improving the illumination robustness.