Facial target detection and keypoints location study using MTCNN model

Abstract

With the extremely rapid development of AI, facial recognition technology in the field of computer vision, which has the advantages of non-contact and easier to collect information, has become a hot topic in all walks of life. Facial target detection and keypoints location are two core parts of facial recognition technology. According to the analysis of the existing facial detection and keypoints location methods, by comparing their advantages and disadvantages, we find the existing problems. Based on the existing research methods, this paper proposes a high-precision facial detection and keypoints location method based on MTCNN model, which highlights the good accuracy of this method compared with the basic CNN or traditional algorithms, so as to improve the accuracy and effectiveness of facial recognition. It has more prominent advantages in this field, and is worthy of scientific research.