KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping

Infrared imaging systems are widely used across industries. However, their output images often exhibit striped noise due to the nonuniform response of the detection system, which significantly affects image quality and visual fidelity. To address challenges such as incomplete stripe removal, potenti...

Full description

Saved in:
Bibliographic Details
Main Authors: Lingxiao Li, Xin Wang, Dan Huang, Yunan He, Zhuqiang Zhong, Qingling Xia
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Photonics Journal
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10978030/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849326838026862592
author Lingxiao Li
Xin Wang
Dan Huang
Yunan He
Zhuqiang Zhong
Qingling Xia
author_facet Lingxiao Li
Xin Wang
Dan Huang
Yunan He
Zhuqiang Zhong
Qingling Xia
author_sort Lingxiao Li
collection DOAJ
description Infrared imaging systems are widely used across industries. However, their output images often exhibit striped noise due to the nonuniform response of the detection system, which significantly affects image quality and visual fidelity. To address challenges such as incomplete stripe removal, potential loss of image details and textures, and the generation of artificial artifacts during destriping, we propose a novel stripe removal method based on a knowledge-embedded diffusion model (KEDM). This approach effectively integrates the spatial distribution characteristics of stripe noise with an innovative, data-driven diffusion network model, creating a hybrid knowledge and data-driven framework for stripe correction. The core components of KEDM are the latent diffusion model (LDM) architecture and the directional wavelet convolution module (DWCM). Specifically, LDM leverages a pretrained variational autoencoder (VAE) to transform the input image into latent feature space for efficient diffusion propagation, reducing computational complexity while preserving image restoration quality. Meanwhile, DWCM uses wavelet convolution operations to construct prior loss functions for stripe noise, precisely guiding the diffusion reconstruction process to achieve a clean, stripe-free image. Empirical evaluations on several benchmark datasets demonstrate that the proposed KEDM outperforms other state-of-the-art destriping algorithms in terms of visual quality and quantitative metrics, validating its excellent performance.
format Article
id doaj-art-7cdb3f45045245a5b5241dc1863bffd4
institution Kabale University
issn 1943-0655
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Photonics Journal
spelling doaj-art-7cdb3f45045245a5b5241dc1863bffd42025-08-20T03:48:02ZengIEEEIEEE Photonics Journal1943-06552025-01-011731910.1109/JPHOT.2025.356483110978030KEDM: Knowledge-Embedded Diffusion Model for Infrared Image DestripingLingxiao Li0https://orcid.org/0000-0001-8491-6814Xin Wang1Dan Huang2Yunan He3Zhuqiang Zhong4https://orcid.org/0000-0002-4293-8246Qingling Xia5School of Science, Chongqing University of Technology, Chongqing, ChinaSchool of Science, Chongqing University of Technology, Chongqing, ChinaChina Research and Development Academy of Machinery Equipment, Beijing, ChinaSchool of Science, Chongqing University of Technology, Chongqing, ChinaSchool of Science, Chongqing University of Technology, Chongqing, ChinaSchool of Artificial Intelligence, Chongqing University of Technology, Chongqing, ChinaInfrared imaging systems are widely used across industries. However, their output images often exhibit striped noise due to the nonuniform response of the detection system, which significantly affects image quality and visual fidelity. To address challenges such as incomplete stripe removal, potential loss of image details and textures, and the generation of artificial artifacts during destriping, we propose a novel stripe removal method based on a knowledge-embedded diffusion model (KEDM). This approach effectively integrates the spatial distribution characteristics of stripe noise with an innovative, data-driven diffusion network model, creating a hybrid knowledge and data-driven framework for stripe correction. The core components of KEDM are the latent diffusion model (LDM) architecture and the directional wavelet convolution module (DWCM). Specifically, LDM leverages a pretrained variational autoencoder (VAE) to transform the input image into latent feature space for efficient diffusion propagation, reducing computational complexity while preserving image restoration quality. Meanwhile, DWCM uses wavelet convolution operations to construct prior loss functions for stripe noise, precisely guiding the diffusion reconstruction process to achieve a clean, stripe-free image. Empirical evaluations on several benchmark datasets demonstrate that the proposed KEDM outperforms other state-of-the-art destriping algorithms in terms of visual quality and quantitative metrics, validating its excellent performance.https://ieeexplore.ieee.org/document/10978030/Conditional diffusion modelinfrared image destripingknowledge-embedded diffusion modelstripe priorwavelet transform
spellingShingle Lingxiao Li
Xin Wang
Dan Huang
Yunan He
Zhuqiang Zhong
Qingling Xia
KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
IEEE Photonics Journal
Conditional diffusion model
infrared image destriping
knowledge-embedded diffusion model
stripe prior
wavelet transform
title KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
title_full KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
title_fullStr KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
title_full_unstemmed KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
title_short KEDM: Knowledge-Embedded Diffusion Model for Infrared Image Destriping
title_sort kedm knowledge embedded diffusion model for infrared image destriping
topic Conditional diffusion model
infrared image destriping
knowledge-embedded diffusion model
stripe prior
wavelet transform
url https://ieeexplore.ieee.org/document/10978030/
work_keys_str_mv AT lingxiaoli kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping
AT xinwang kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping
AT danhuang kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping
AT yunanhe kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping
AT zhuqiangzhong kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping
AT qinglingxia kedmknowledgeembeddeddiffusionmodelforinfraredimagedestriping