Abstract:Hairstyle transfer aims to combine the hairstyle attributes of the reference images with the face and background of the identity image into one realistic image. This is challenging due to the differences in semantic structure between reference images and identity image, as well as the tangled hairstyle attributes. Existing GAN-based methods tend to simply combine image patches of hair, face, and background, which leads to discordant semantic structures and obvious artifacts in the results. And the GAN inversion-based methods require elaborate orthogonalization of loss gradients with additional calculations to alleviate the entanglement between the complex hairstyle attributes. To address these limitations, we propose a novel framework for hairstyle transfer based on StyleGAN inversion. Specifically, instead of simply combining the image patches, we use StyleGAN’s generative power to get more reasonable target semantic masks. During the alignment with the target masks, background inpainting is performed by StyleGAN rather than auxiliary models, which makes the results more realistic. Furthermore, our method decomposes hairstyle into three attributes: structure, texture, and color to better control the hairstyle transfer process. Besides, we observe that the spatial scales of these attributes match those of the StyleGAN’s latent codes. Then, we decouple the latent code in the hierarchical dimension and use different layers of the latent code to complete the transfer of different hair attributes, which reduces the entanglement between attributes without extra calculations. Correspondingly, our method divides hairstyle transfer into three steps, making the transfer process more controllable. By comparing with state-of-the-art methods, we verify the effectiveness of our method.

HairGAN: Spatial-Aware Palette GAN for Hair Color Transfer

Ihair Recolorer: Deep Image-to-video Hair Color Transfer

Hair-GAN: Recovering 3D Hair Structure from a Single Image Using Generative Adversarial Networks

HairFastGAN: Realistic and Robust Hair Transfer with a Fast Encoder-Based Approach

CFMNet - Coarse-to-Fine Cascaded Feature Mapping Network for Hair Attribute Transfer.

HairMapper: Removing Hair from Portraits Using GANs

Hair-GANs: Recovering 3D Hair Structure from a Single Image

Hairstyle transfer via manipulating decoupled latent codes of StyleGAN2

HSEGAN: Hair Synthesis and Editing Using Structure-Adaptive Normalization on Generative Adversarial Network

Personalized hairstyle and hair color editing based on multi-feature fusion

PSGAN: Pose and Expression Robust Spatial-Aware GAN for Customizable Makeup Transfer

Efficient Hair Style Transfer with Generative Adversarial Networks

Makeup Transfer Based on Generative Adversarial Network for Large Angle Spatial Misalignment.

StyleGAN Salon: Multi-View Latent Optimization for Pose-Invariant Hairstyle Transfer

PSGAN: Pose-Robust Spatial-Aware GAN for Customizable Makeup Transfer.

PSGAN: Pose-Robust Spatial-Aware GAN for Customizable Makeup Transfer

Two-phase Hair Image Synthesis by Self-Enhancing Generative Model

Shape-preserving Swapping Autoencoder for Garment Texture Transfer

Spatial Content Alignment For Pose Transfer

Composite makeup transfer model based on generative adversarial networks

A robust transformer GAN for unpaired data makeup transfer