BFS: Back-to-Front Layered Image Synthesis via Knowledge Transfer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kang, Kyoungkook, Sim, Gyujin, Cho, Sunghyun
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911713894334464
author Kang, Kyoungkook
Sim, Gyujin
Cho, Sunghyun
author_facet Kang, Kyoungkook
Sim, Gyujin
Cho, Sunghyun
contents As generative models expand the possibilities of visual content creation, layered image synthesis has emerged as a promising direction for controllable and creative editing. However, existing methods struggle to fully realize this potential. Decomposition-based methods often struggle with clean separation, while generation-based methods suffer from difficulty in training data acquisition, reducing quality and scene diversity. In this paper, we propose BFS, a novel generation-based framework for layered image synthesis. Specifically, given a background image and user guidance, BFS synthesizes a foreground layer that incorporates not only a foreground object but also its associated visual effects, such as shadows and reflections, while seamlessly harmonizing with the background to produce a coherent composite. To enable diverse and high-quality foreground layer synthesis while overcoming data scarcity, we leverage the comparatively easy-to-learn knowledge of unlayered image synthesis for the foreground synthesis. To this end, we adopt a dual-branch diffusion framework in which two interconnected branches generate a composite image and a foreground layer, respectively, enabling bidirectional knowledge transfer. Based on this framework, we propose a two-stage training scheme that utilizes a high-quality unlayered composite image dataset to effectively enhance foreground quality. Extensive experiments, including a user study, show that BFS produces high-quality layered images, consistently outperforming prior methods.
format Preprint
id arxiv_https___arxiv_org_abs_2605_24894
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle BFS: Back-to-Front Layered Image Synthesis via Knowledge Transfer
Kang, Kyoungkook
Sim, Gyujin
Cho, Sunghyun
Computer Vision and Pattern Recognition
As generative models expand the possibilities of visual content creation, layered image synthesis has emerged as a promising direction for controllable and creative editing. However, existing methods struggle to fully realize this potential. Decomposition-based methods often struggle with clean separation, while generation-based methods suffer from difficulty in training data acquisition, reducing quality and scene diversity. In this paper, we propose BFS, a novel generation-based framework for layered image synthesis. Specifically, given a background image and user guidance, BFS synthesizes a foreground layer that incorporates not only a foreground object but also its associated visual effects, such as shadows and reflections, while seamlessly harmonizing with the background to produce a coherent composite. To enable diverse and high-quality foreground layer synthesis while overcoming data scarcity, we leverage the comparatively easy-to-learn knowledge of unlayered image synthesis for the foreground synthesis. To this end, we adopt a dual-branch diffusion framework in which two interconnected branches generate a composite image and a foreground layer, respectively, enabling bidirectional knowledge transfer. Based on this framework, we propose a two-stage training scheme that utilizes a high-quality unlayered composite image dataset to effectively enhance foreground quality. Extensive experiments, including a user study, show that BFS produces high-quality layered images, consistently outperforming prior methods.
title BFS: Back-to-Front Layered Image Synthesis via Knowledge Transfer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.24894