DCT-Net: Domain-Calibrated Translation for Portrait Stylization

by   Yifang Men, et al.

This paper introduces DCT-Net, a novel image translation architecture for few-shot portrait stylization. Given limited style exemplars (∼100), the new architecture can produce high-quality style transfer results with advanced ability to synthesize high-fidelity contents and strong generality to handle complicated scenes (e.g., occlusions and accessories). Moreover, it enables full-body image translation via one elegant evaluation network trained by partial observations (i.e., stylized heads). Few-shot learning based style transfer is challenging since the learned model can easily become overfitted in the target domain, due to the biased distribution formed by only a few training examples. This paper aims to handle the challenge by adopting the key idea of "calibration first, translation later" and exploring the augmented global structure with locally-focused translation. Specifically, the proposed DCT-Net consists of three modules: a content adapter borrowing the powerful prior from source photos to calibrate the content distribution of target samples; a geometry expansion module using affine transformations to release spatially semantic constraints; and a texture translation module leveraging samples produced by the calibrated distribution to learn a fine-grained conversion. Experimental results demonstrate the proposed method's superiority over the state of the art in head stylization and its effectiveness on full image translation with adaptive deformations.


page 1

page 6

page 7

page 8

page 9


Diffusion-based Image Translation using Disentangled Style and Content Representation

Diffusion-based image translation guided by semantic texts or a single t...

Training-free Style Transfer Emerges from h-space in Diffusion models

Diffusion models (DMs) synthesize high-quality images in various domains...

Global and Local Alignment Networks for Unpaired Image-to-Image Translation

The goal of unpaired image-to-image translation is to produce an output ...

XGAN: Unsupervised Image-to-Image Translation for Many-to-Many Mappings

Style transfer usually refers to the task of applying color and texture ...

STALP: Style Transfer with Auxiliary Limited Pairing

We present an approach to example-based stylization of images that uses ...

Few-shot Face Image Translation via GAN Prior Distillation

Face image translation has made notable progress in recent years. Howeve...

Free Lunch for Few-shot Learning: Distribution Calibration

Learning from a limited number of samples is challenging since the learn...

Please sign up or login with your details

Forgot password? Click here to reset