Multiview Face Automatic Recognition Method Based on Geometric Characteristics

Abstract

At present, most face recognition methods are aimed at frontal faces, but there are still great difficulties in multiview face detection, and there are few effective methods. In this paper, considering the multiview situation in face recognition, a multiview face automatic recognition method based on geometric characteristics is proposed. First, multiview face image preprocessing is performed, and then the skin color model is established. After processing, the elliptical area of the face is determined to realize face detection. Then, the geometric characteristics of the multiview face are extracted, and finally, the minimum distance classifier is established. According to the extracted multiview people face geometry features, complete multiview face automatic recognition is performed. The results show that compared with knowledge-based methods, template matching-based methods, and statistical learning methods, the research method has higher recognition accuracy and shorter running time, indicating that the research method is simple and easy to implement and is more suitable for real-time needs. Researched face recognition technology.