Color Alignment in Diffusion

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shum, Ka Chun, Hua, Binh-Son, Nguyen, Duc Thanh, Yeung, Sai-Kit
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917949846061056
author Shum, Ka Chun
Hua, Binh-Son
Nguyen, Duc Thanh
Yeung, Sai-Kit
author_facet Shum, Ka Chun
Hua, Binh-Son
Nguyen, Duc Thanh
Yeung, Sai-Kit
contents Diffusion models have shown great promise in synthesizing visually appealing images. However, it remains challenging to condition the synthesis at a fine-grained level, for instance, synthesizing image pixels following some generic color pattern. Existing image synthesis methods often produce contents that fall outside the desired pixel conditions. To address this, we introduce a novel color alignment algorithm that confines the generative process in diffusion models within a given color pattern. Specifically, we project diffusion terms, either imagery samples or latent representations, into a conditional color space to align with the input color distribution. This strategy simplifies the prediction in diffusion models within a color manifold while still allowing plausible structures in generated contents, thus enabling the generation of diverse contents that comply with the target color pattern. Experimental results demonstrate our state-of-the-art performance in conditioning and controlling of color pixels, while maintaining on-par generation quality and diversity in comparison with regular diffusion models.
format Preprint
id arxiv_https___arxiv_org_abs_2503_06746
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Color Alignment in Diffusion
Shum, Ka Chun
Hua, Binh-Son
Nguyen, Duc Thanh
Yeung, Sai-Kit
Computer Vision and Pattern Recognition
Diffusion models have shown great promise in synthesizing visually appealing images. However, it remains challenging to condition the synthesis at a fine-grained level, for instance, synthesizing image pixels following some generic color pattern. Existing image synthesis methods often produce contents that fall outside the desired pixel conditions. To address this, we introduce a novel color alignment algorithm that confines the generative process in diffusion models within a given color pattern. Specifically, we project diffusion terms, either imagery samples or latent representations, into a conditional color space to align with the input color distribution. This strategy simplifies the prediction in diffusion models within a color manifold while still allowing plausible structures in generated contents, thus enabling the generation of diverse contents that comply with the target color pattern. Experimental results demonstrate our state-of-the-art performance in conditioning and controlling of color pixels, while maintaining on-par generation quality and diversity in comparison with regular diffusion models.
title Color Alignment in Diffusion
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.06746