Face Shape Classification Based on MTCNN and FaceNet

Deep learning has made great progress in the field of face recognition. Most face shape classification methods use the distance and Angle between facial features and the landmarks of facial outline, and calculate some values through formulas. With the development of deep learning, many methods take face shape as the learning object. After comparing some networks, this paper uses MTCNN and FaceNet, and the accuracy is 92.32%, which is higher than other networks. In addition, Center Loss is used instead of Triplet Loss to achieve a better effect.