Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Jingxi, Zhang, Yixiao, Qian, Xiaoye, Li, Zongxia, Fermuller, Cornelia, Chen, Caren, Aloimonos, Yiannis
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2511.20996
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911538434015232
author Chen, Jingxi
Zhang, Yixiao
Qian, Xiaoye
Li, Zongxia
Fermuller, Cornelia
Chen, Caren
Aloimonos, Yiannis
author_facet Chen, Jingxi
Zhang, Yixiao
Qian, Xiaoye
Li, Zongxia
Fermuller, Cornelia
Chen, Caren
Aloimonos, Yiannis
contents Images can be viewed as layered compositions, foreground objects over background, with potential occlusions. This layered representation enables independent editing of elements, offering greater flexibility for content creation. Despite the progress in large generative models, decomposing a single image into layers remains challenging due to limited methods and data. We observe a strong connection between layer decomposition and in/outpainting tasks, and propose adapting a diffusion-based inpainting model for layer decomposition using lightweight finetuning. To further preserve detail in the latent space, we introduce a novel multi-modal context fusion module with linear attention complexity. Our model is trained purely on a synthetic dataset constructed from open-source assets and achieves superior performance in object removal and occlusion recovery, unlocking new possibilities in downstream editing and creative applications.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20996
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle From Inpainting to Layer Decomposition: Repurposing Generative Inpainting Models for Image Layer Decomposition
Chen, Jingxi
Zhang, Yixiao
Qian, Xiaoye
Li, Zongxia
Fermuller, Cornelia
Chen, Caren
Aloimonos, Yiannis
Computer Vision and Pattern Recognition
Images can be viewed as layered compositions, foreground objects over background, with potential occlusions. This layered representation enables independent editing of elements, offering greater flexibility for content creation. Despite the progress in large generative models, decomposing a single image into layers remains challenging due to limited methods and data. We observe a strong connection between layer decomposition and in/outpainting tasks, and propose adapting a diffusion-based inpainting model for layer decomposition using lightweight finetuning. To further preserve detail in the latent space, we introduce a novel multi-modal context fusion module with linear attention complexity. Our model is trained purely on a synthetic dataset constructed from open-source assets and achieves superior performance in object removal and occlusion recovery, unlocking new possibilities in downstream editing and creative applications.
title From Inpainting to Layer Decomposition: Repurposing Generative Inpainting Models for Image Layer Decomposition
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.20996