Scribble-Guided Diffusion for Training-free Text-to-Image Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Seonho, Choi, Jiho, Lim, Seohyun, Kim, Jiwook, Shim, Hyunjung
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916390952239104
author Lee, Seonho
Choi, Jiho
Lim, Seohyun
Kim, Jiwook
Shim, Hyunjung
author_facet Lee, Seonho
Choi, Jiho
Lim, Seohyun
Kim, Jiwook
Shim, Hyunjung
contents Recent advancements in text-to-image diffusion models have demonstrated remarkable success, yet they often struggle to fully capture the user's intent. Existing approaches using textual inputs combined with bounding boxes or region masks fall short in providing precise spatial guidance, often leading to misaligned or unintended object orientation. To address these limitations, we propose Scribble-Guided Diffusion (ScribbleDiff), a training-free approach that utilizes simple user-provided scribbles as visual prompts to guide image generation. However, incorporating scribbles into diffusion models presents challenges due to their sparse and thin nature, making it difficult to ensure accurate orientation alignment. To overcome these challenges, we introduce moment alignment and scribble propagation, which allow for more effective and flexible alignment between generated images and scribble inputs. Experimental results on the PASCAL-Scribble dataset demonstrate significant improvements in spatial control and consistency, showcasing the effectiveness of scribble-based guidance in diffusion models. Our code is available at https://github.com/kaist-cvml-lab/scribble-diffusion.
format Preprint
id arxiv_https___arxiv_org_abs_2409_08026
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Scribble-Guided Diffusion for Training-free Text-to-Image Generation
Lee, Seonho
Choi, Jiho
Lim, Seohyun
Kim, Jiwook
Shim, Hyunjung
Computer Vision and Pattern Recognition
Recent advancements in text-to-image diffusion models have demonstrated remarkable success, yet they often struggle to fully capture the user's intent. Existing approaches using textual inputs combined with bounding boxes or region masks fall short in providing precise spatial guidance, often leading to misaligned or unintended object orientation. To address these limitations, we propose Scribble-Guided Diffusion (ScribbleDiff), a training-free approach that utilizes simple user-provided scribbles as visual prompts to guide image generation. However, incorporating scribbles into diffusion models presents challenges due to their sparse and thin nature, making it difficult to ensure accurate orientation alignment. To overcome these challenges, we introduce moment alignment and scribble propagation, which allow for more effective and flexible alignment between generated images and scribble inputs. Experimental results on the PASCAL-Scribble dataset demonstrate significant improvements in spatial control and consistency, showcasing the effectiveness of scribble-based guidance in diffusion models. Our code is available at https://github.com/kaist-cvml-lab/scribble-diffusion.
title Scribble-Guided Diffusion for Training-free Text-to-Image Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.08026