Overview of Mainstream Face Recognition Methods

Abstract

Human face is one of the most important features of human beings. In the information age, the face has personal information and emotional expression, social communication, public safety and other extremely important attributes. Due to the characteristics of inconspicuous facial features and small differences, traditional face recognition by manual training leads to low efficiency and high error rate. Based on the advantages of computer automation, face recognition technology has tended to be efficient, accurate and intelligent. Face recognition in a broad sense includes four stages: portrait image preprocessing, eigenface display, data training and data production. The use of CNN neural network can deal with complex abstract facial features. OpenCV open-source function libraries and Python simple and efficient machine language have advantages. This paper will focus on CNN deep neural network face recognition principle and OpenCV face recognition process two mainstream methods, as well as their defects comparison and conclusions.