Enregistré dans:
Détails bibliographiques
Auteurs principaux: Mayet, Tsiry, Shamsolmoali, Pourya, Bernard, Simon, Granger, Eric, Hérault, Romain, Chatelain, Clement
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2410.09306
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915356969271296
author Mayet, Tsiry
Shamsolmoali, Pourya
Bernard, Simon
Granger, Eric
Hérault, Romain
Chatelain, Clement
author_facet Mayet, Tsiry
Shamsolmoali, Pourya
Bernard, Simon
Granger, Eric
Hérault, Romain
Chatelain, Clement
contents Diffusion models have emerged as highly effective techniques for inpainting, however, they remain constrained by slow sampling rates. While recent advances have enhanced generation quality, they have also increased sampling time, thereby limiting scalability in real-world applications. We investigate the generative sampling process of diffusion-based inpainting models and observe that these models make minimal use of the input condition during the initial sampling steps. As a result, the sampling trajectory deviates from the data manifold, requiring complex synchronization mechanisms to realign the generation process. To address this, we propose Time-aware Diffusion Paint (TD-Paint), a novel approach that adapts the diffusion process by modeling variable noise levels at the pixel level. This technique allows the model to efficiently use known pixel values from the start, guiding the generation process toward the target manifold. By embedding this information early in the diffusion process, TD-Paint significantly accelerates sampling without compromising image quality. Unlike conventional diffusion-based inpainting models, which require a dedicated architecture or an expensive generation loop, TD-Paint achieves faster sampling times without architectural modifications. Experimental results across three datasets show that TD-Paint outperforms state-of-the-art diffusion models while maintaining lower complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2410_09306
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TD-Paint: Faster Diffusion Inpainting Through Time Aware Pixel Conditioning
Mayet, Tsiry
Shamsolmoali, Pourya
Bernard, Simon
Granger, Eric
Hérault, Romain
Chatelain, Clement
Computer Vision and Pattern Recognition
Diffusion models have emerged as highly effective techniques for inpainting, however, they remain constrained by slow sampling rates. While recent advances have enhanced generation quality, they have also increased sampling time, thereby limiting scalability in real-world applications. We investigate the generative sampling process of diffusion-based inpainting models and observe that these models make minimal use of the input condition during the initial sampling steps. As a result, the sampling trajectory deviates from the data manifold, requiring complex synchronization mechanisms to realign the generation process. To address this, we propose Time-aware Diffusion Paint (TD-Paint), a novel approach that adapts the diffusion process by modeling variable noise levels at the pixel level. This technique allows the model to efficiently use known pixel values from the start, guiding the generation process toward the target manifold. By embedding this information early in the diffusion process, TD-Paint significantly accelerates sampling without compromising image quality. Unlike conventional diffusion-based inpainting models, which require a dedicated architecture or an expensive generation loop, TD-Paint achieves faster sampling times without architectural modifications. Experimental results across three datasets show that TD-Paint outperforms state-of-the-art diffusion models while maintaining lower complexity.
title TD-Paint: Faster Diffusion Inpainting Through Time Aware Pixel Conditioning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.09306