Face restoration method based on multi-channel attention selection generative adversarial network

Abstract

The invention provides a face restoration method based on a multi-channel attention selection generative adversarial network. The method comprises the following steps of S1, collecting face data and performing preprocessing; s2, establishing a face restoration model and a loss function; s3, in the first stage, learning the image to generate a subnet Gi and preliminarily repairing the image; s4, inthe second stage, an intermediate output image IG is generated, and a multi-channel attention image IA is learned; s5, constructing a multi-channel attention selection model and outputting a final composite graph; s6, carrying out face restoration. Wherein the face restoration model comprises a generator network Gi, a parameter sharing discriminator D and a multi-channel attention selection network Ga, and the loss functions comprise an uncertainty pixel loss function and a resistance loss function. According to the face restoration method provided by the invention, the uncertainty graph is effectively learned to guide pixel loss, so that stronger optimization is realized, and a better face restoration method is provided.

Description

【Technical field】 [0001] The invention relates to the fields of deep learning and image processing, in particular to a face restoration method based on a multi-channel attention selection generative confrontation network. 【Background technique】 [0002] In the field of image inpainting, especially for intraocular painting, most deep learning techniques fail to preserve the identity of people in photos, although DNNs (deep neural networks) can produce semantically plausible and realistic-looking results. For example, a DNN can learn to open a pair of closed eyes, but the DNN itself cannot guarantee that the new eyes will correspond to the specific ocular structure of hominids. [0003] GAN (Generative adversarial networks) is a specific type of deep network that includes a learnable adversarial loss function represented by a discriminator network. GANs have been successfully used to generate faces from scratch, or to map missing regions on faces, suitable for general face ma...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the problems of image occlusion, incorrect coloring repair and strange artificial repair traces in face repair technology under certain conditions, the present invention provides a face repair method based on multi-channel attention selection generative confrontation network

Images