Virtual makeup removing method for simulating makeup image based on dual generative adversarial network

Abstract

The invention provides a virtual makeup removal method for simulating a makeup image based on a dual generative adversarial network, and solves the technical problem that the makeup removal effect is poor due to the fact that simulated makeup is influenced by facial information of a simulator and a simulated person. The method comprises the following steps: firstly, combining a makeup face image of a simulator and a face image of a simulation object into a first input matrix, and inputting the first input matrix into a makeup removal generator network to generate an image of the imitator without markup; secondly, combining the image of the imitator without makeup and the face image of the imitating object into a second input matrix, and inputting the second input matrix into a makeup generator network to generate a makeup image; then, respectively inputting the image without makeup and the makeup image into a discriminator network, calculating a loss function of the discriminator, and adjusting network parameters according to the loss function to obtain a makeup removal generator; and finally, inputting a to-be-detected face image into the makeup removal generator to obtain an image of the user without makeup. According to the method, the makeup image and the simulated person image are input into the makeup removal generator at the same time, and the good makeup removing effect is achieved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a virtual makeup removal method for imitating makeup images based on dual generative confrontation networks. Background technique [0002] Everyone has a heart for beauty. With the improvement of people's quality of life, more and more people pay attention to their appearance, hoping to improve their image through makeup. But while makeup gives people confidence, it also brings some inconvenience to people's lives, including the decline in the recognition performance of the face authentication system. [0003] In recent years, the development of deep learning technology has improved the recognition rate of face authentication systems by leaps and bounds. Therefore, face recognition technology has been used more and more in actual scenarios such as human-computer interaction, credit card verification, and access control systems. For example, at the opening ceremony of t...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the deficiencies in the above-mentioned background technology, the present invention proposes a virtual makeup removal method for imitating makeup images based on dual generative adversarial networks, which solves the problem that existing simulated makeup is affected by the facial information of the imitator and the imitated, resulting in a makeup removal effect. poor technical problem

Images