Face Attribute Invertion

TL;DR

This paper introduces a novel GAN-based method for automatic face attribute inversion using a single generator, emphasizing stability and detail preservation through a multi-loss strategy and a modified U-net structure.

Contribution

The paper presents a new self-perception approach that simplifies face attribute inversion by eliminating the need for multiple generators or conditional inputs.

Findings

Stable training achieved with the proposed model

Finer facial details are preserved during inversion

Single generator approach reduces complexity

Abstract

Manipulating human facial images between two domains is an important and interesting problem. Most of the existing methods address this issue by applying two generators or one generator with extra conditional inputs. In this paper, we proposed a novel self-perception method based on GANs for automatical face attribute inverse. The proposed method takes face images as inputs and employs only one single generator without being conditioned on other inputs. Profiting from the multi-loss strategy and modified U-net structure, our model is quite stable in training and capable of preserving finer details of the original face images.