Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Face attribute conversion method through neural network

A neural network and attribute technology, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve problems that consume a lot of time and energy, and pictures are not human faces

Active Publication Date: 2018-02-06
SUN YAT SEN UNIV
View PDF3 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] The face attribute is the description of the facial image. Usually, people use PS manual modification to achieve the effect of face attribute transfer. It takes a lot of time and energy to get a satisfactory result.
In addition, the method of using attribute discriminant network backpropagation to modify the original image can also achieve the purpose of face attribute transfer, but the image obtained by this method may not be a face

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Face attribute conversion method through neural network
  • Face attribute conversion method through neural network
  • Face attribute conversion method through neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0035] A method for realizing face attribute conversion through a neural network in the present invention needs to train two networks, one is the generation network G-Net in GAN (generative confrontation network), and the other is the attribute discrimination network E-Net. Among them, G-Net is responsible for generating images, that is, inputting a random vector can obtain a visually realistic face image. E-Net is responsible for identifying attributes, that is, judging whether the current picture has the attributes we have defined. G-Net and E-Net are trained using real face images.

[0036] The training method of G-Net. The training process of G-Net needs to be equipped with a discriminant network D-Net, but after G-Net completes the training, D-Net no longer needs to be used. See attached picture for training structure diagram figure 1 . The positioning of G-Net is to generate images, and the positioning of D-Net is to distinguish as much as possible whether the image ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a face attribute conversion method through the neural network. The generation network G-Net is trained, and the generation network G-Net is used for generating images, namely acquiring a visually real face image through inputting a random vector; an attribute discrimination network E-Net is trained, and the attribute discrimination network E-Net is used for discriminating attributes, namely discriminating whether a present picture has limited attributes; after the generation network G-Net and the attribute discrimination network E-Net are trained, the generation networkG-Net and the attribute discrimination network E-Net are connected in series, output of the G-Net is input of the E-Net, and face attribute conversion operation is carried out. The method is advantaged in that pictures with the natural effect can be rapidly generated, a problem that the generation results are non-natural faces or non-human faces is solved, and secondary manual modification is notneeded.

Description

technical field [0001] The invention relates to the field of digital image processing, and more specifically, relates to a method for realizing conversion of human face attributes through a neural network. Background technique [0002] Face attributes are the description of facial images. Usually, people use PS to manually modify to achieve the effect of face attribute transfer. It takes a lot of time and energy to get a satisfactory result. In addition, the method of modifying the original image by using attribute discriminant network backpropagation can also achieve the purpose of face attribute transfer, but the image obtained by this method may not be a human face. Contents of the invention [0003] The invention provides a method for realizing human face attribute conversion through a neural network, and the method can quickly generate pictures with natural effects. [0004] In order to achieve the above-mentioned technical effect, the technical scheme of the present...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06N3/04G06N3/08
CPCG06N3/084G06V40/16G06N3/045
Inventor 孔方圆丁圣勇朝红阳
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com