A Face Recognition Method Based on Deep Learning in Real Scenes

Abstract

The present invention provides a face recognition method based on deep learning in a real scene, which is proposed to solve the problem that the existing face recognition method in a real scene can only solve the influence of a single factor, but cannot solve the influence of factors such as posture and illumination , including: using an existing face detector to predict the face position of each picture in the training database, and intercepting and saving the real face and non-face images; according to down-sampling of face images and non-face images Corresponding low-resolution images; build a generative confrontation network, the generative confrontation network includes a generator and a discriminator; the generator further includes an upsampling network and an optimization network; use high-resolution face, non-face images and corresponding low-resolution Face and non-face images are used to train the generative confrontation network; the position of the face is marked in the input picture according to the score of the discriminator on the face candidate area obtained from the existing face detector. The invention is applicable to the recognition and detection of human faces.

Description

technical field [0001] The invention relates to the field of face recognition, in particular to a face recognition method based on deep learning in a real scene. Background technique [0002] With the development of e-commerce and other applications, face recognition has become the most potential biometric authentication method. This application background requires automatic face recognition systems to have certain recognition capabilities for face images in real scenes. Therefore, Faced with a series of problems, face detection began to be regarded as an independent subject by researchers. In addition, face recognition technology in real scenes has urgent application needs in many fields such as security, criminal investigation, search and rescue. [0003] Based on the fact that face detection has very important basic research value and urgent application requirements in the field of machine vision, the corresponding technologies for face detection are also constantly bein...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the face recognition method under the existing real scene can only solve the influence of a single factor, but can not solve the influence of other influencing factors such as posture, illumination; And the face recognition method based on the face alignment method can alleviate The influence of posture, but it cannot solve the shortcomings of other factors such as scale and blur, and proposes a face recognition method based on deep learning in real scenes, including:

Images