Tell Me What You See: Text-Guided Real-World Image Denoising

Image reconstruction from noisy sensor measurements is challenging and many methods have been proposed for it. Yet, most approaches focus on learning robust natural image priors while modeling the scene’s noise statistics. In extremely low-light conditions, these methods often remain insu...

詳細記述

保存先:
書誌詳細
主要な著者: Erez Yosef, Raja Giryes
フォーマット: 論文
言語:英語
出版事項: IEEE 2025-01-01
シリーズ:IEEE Open Journal of Signal Processing
主題:
オンライン・アクセス:https://ieeexplore.ieee.org/document/11078899/
タグ: タグ追加
タグなし, このレコードへの初めてのタグを付けませんか!
_version_ 1839609170210324480
author Erez Yosef
Raja Giryes
author_facet Erez Yosef
Raja Giryes
author_sort Erez Yosef
collection DOAJ
description Image reconstruction from noisy sensor measurements is challenging and many methods have been proposed for it. Yet, most approaches focus on learning robust natural image priors while modeling the scene’s noise statistics. In extremely low-light conditions, these methods often remain insufficient. Additional information is needed, such as multiple captures or, as suggested here, scene description. As an alternative, we propose using a text-based description of the scene as an additional prior, something the photographer can easily provide. Inspired by the remarkable success of text-guided diffusion models in image generation, we show that adding image caption information significantly improves image denoising and reconstruction for both synthetic and real-world images. All code and data will be made publicly available upon publication.
format Article
id doaj-art-ecccd4bcaf944b40a83a33aca09dfefb
institution Matheson Library
issn 2644-1322
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Open Journal of Signal Processing
spelling doaj-art-ecccd4bcaf944b40a83a33aca09dfefb2025-07-30T23:00:36ZengIEEEIEEE Open Journal of Signal Processing2644-13222025-01-01689089910.1109/OJSP.2025.358871511078899Tell Me What You See: Text-Guided Real-World Image DenoisingErez Yosef0https://orcid.org/0000-0003-1503-189XRaja Giryes1https://orcid.org/0000-0002-2830-0297Tel-Aviv University, Tel Aviv, IsraelTel-Aviv University, Tel Aviv, IsraelImage reconstruction from noisy sensor measurements is challenging and many methods have been proposed for it. Yet, most approaches focus on learning robust natural image priors while modeling the scene’s noise statistics. In extremely low-light conditions, these methods often remain insufficient. Additional information is needed, such as multiple captures or, as suggested here, scene description. As an alternative, we propose using a text-based description of the scene as an additional prior, something the photographer can easily provide. Inspired by the remarkable success of text-guided diffusion models in image generation, we show that adding image caption information significantly improves image denoising and reconstruction for both synthetic and real-world images. All code and data will be made publicly available upon publication.https://ieeexplore.ieee.org/document/11078899/Computational imagingdeep learningartificial intelligence
spellingShingle Erez Yosef
Raja Giryes
Tell Me What You See: Text-Guided Real-World Image Denoising
IEEE Open Journal of Signal Processing
Computational imaging
deep learning
artificial intelligence
title Tell Me What You See: Text-Guided Real-World Image Denoising
title_full Tell Me What You See: Text-Guided Real-World Image Denoising
title_fullStr Tell Me What You See: Text-Guided Real-World Image Denoising
title_full_unstemmed Tell Me What You See: Text-Guided Real-World Image Denoising
title_short Tell Me What You See: Text-Guided Real-World Image Denoising
title_sort tell me what you see text guided real world image denoising
topic Computational imaging
deep learning
artificial intelligence
url https://ieeexplore.ieee.org/document/11078899/
work_keys_str_mv AT erezyosef tellmewhatyouseetextguidedrealworldimagedenoising
AT rajagiryes tellmewhatyouseetextguidedrealworldimagedenoising