DyStyle: Dynamic Neural Network for Multi-Attribute-Conditioned Style Editing

by   Bingchuan Li, et al.

Great diversity and photorealism have been achieved by unconditional GAN frameworks such as StyleGAN and its variations. In the meantime, persistent efforts have been made to enhance the semantic controllability of StyleGANs. For example, a dozen of style manipulation methods have been recently proposed to perform attribute-conditioned style editing. Although some of these methods work well in manipulating the style codes along one attribute, the control accuracy when jointly manipulating multiple attributes tends to be problematic. To address these limitations, we propose a Dynamic Style Manipulation Network (DyStyle) whose structure and parameters vary by input samples, to perform nonlinear and adaptive manipulation of latent codes for flexible and precise attribute control. Additionally, a novel easy-to-hard training procedure is introduced for efficient and stable training of the DyStyle network. Extensive experiments have been conducted on faces and other objects. As a result, our approach demonstrates fine-grained disentangled edits along multiple numeric and binary attributes. Qualitative and quantitative comparisons with existing style manipulation methods verify the superiority of our method in terms of the attribute control accuracy and identity preservation without compromising the photorealism. The advantage of our method is even more significant for joint multi-attribute control. The source codes are made publicly available at \href{https://github.com/phycvgan/DyStyle}{phycvgan/DyStyle}.


page 14

page 15

page 16

page 17

page 18

page 19

page 20

page 21


Everything is There in Latent Space: Attribute Editing and Attribute Style Manipulation by StyleGAN Latent Space Exploration

Unconstrained Image generation with high realism is now possible using r...

StyleFlow: Attribute-conditioned Exploration of StyleGAN-Generated Images using Conditional Continuous Normalizing Flows

High-quality, diverse, and photorealistic images can now be generated by...

Supervised Attribute Information Removal and Reconstruction for Image Manipulation

The goal of attribute manipulation is to control specified attribute(s) ...

FaceController: Controllable Attribute Editing for Face in the Wild

Face attribute editing aims to generate faces with one or multiple desir...

SMILE: Semantically-guided Multi-attribute Image and Layout Editing

Attribute image manipulation has been a very active topic since the intr...

SAM: Semantic Attribute Modulation for Language Modeling and Style Variation

This paper presents a Semantic Attribute Modulation (SAM) for language m...

MaskGAN: Towards Diverse and Interactive Facial Image Manipulation

Facial image manipulation has achieved great progresses in recent years....

Please sign up or login with your details

Forgot password? Click here to reset