Face aging image synthesis method based on cyclic conditional generative adversarial network

An image synthesis and conditional generation technology, applied in the field of computer vision, can solve the problems of lack of aging methods for aging images, failure to maintain identity consistency, and poor quality of aging image generation

CN111798369AActive Publication Date: 2020-10-20UNIV OF ELECTRONIC SCI & TECH OF CHINA
4 Cites 9 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2020-10-20

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a face aging image synthesis method based on a cyclic conditional generative adversarial network, and belongs to the field of computer vision. The method comprises the steps offirstly selecting a generative adversarial network as a basic framework; meanwhile, taking the idea of dual learning of the cyclic generative adversarial network as reference; utilizing a supervisedlearning thought of an auxiliary discriminator; introducing a category label innovatively when the cyclic generative adversarial network is used for generating an aging picture; according to the method, the network is enabled to increase attention to specific age characteristics, an auxiliary classification branch is added to the discriminator so that the generation network is enabled to effectively utilize label information to learn specific knowledge, and generation and conversion of the generation network in images of different age groups can be completed through single training via the idea of dual learning. Through the method, the advantages of dual learning and auxiliary classification supervision thought are fully utilized, and the efficiency and picture quality of the cyclic generative adversarial network in senescence image generation are greatly improved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the field of computer vision, and mainly relates to the problem of synthesizing human face aging images. It is mainly used in the detection of criminal cases and the film and television entertainment industry. Background technique

[0002] Face aging image synthesis is the use of computer vision related technologies to synthesize corresponding images of specified face images in different age groups on the basis of retaining the identity characteristics of the original face image. With the continuous breakthroughs in relevant theoretical research and the rapid development of the information technology industry, the research on face aging image synthesis is very extensive in public security criminal investigation, cross-aging face recognition, face data translation, and face data set expansion. Applications. In recent years, face aging image synthesis has been further developed based on the development of machine learning and d...

Examples

Embodiment Construction

[0061] This method is based on the cyclic condition to generate a face aging image synthesis method against the network, and its realization includes the following steps:

[0062] Step 1: Preprocess the dataset. Use the UTKFace (https: / / susanqq.github.io / UTKFace / ) face database for face alignment. The database picture first needs to calculate the position and size of the face in the image, which is represented by a square box. Secondly, use the box to cut out the face part from the whole image as the input for the subsequent steps. Then locate the coordinates of the feature points in the picture, here a positioning algorithm of 68 feature points is used. Finally, according to the located feature point coordinates, the geometric mean point of the face and the geometric center coordinates of the left and right eyes are calculated. Then, according to these information, calculate the parameters of the rotation, translation and scaling transformation that the image needs to perf...