Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index.

Zero-Shot Medical Image Translation via Frequency-Guided Diffusion Models

Open access
Li, Yunxiang, Shao, Hua-Chieh, Liang, Xiao, Chen, Liyuan, Li, Ruiqi, Jiang, Steve, Wang, Jing, Zhang, You · 2024 · IEEE Transactions on Medical Imaging

Abstract

Recently, the diffusion model has emerged as a superior generative model that can produce high quality and realistic images. However, for medical image translation, the existing diffusion models are deficient in accurately retaining structural information since the structure details of source domain images are lost during the forward diffusion process and cannot be fully recovered through learned reverse diffusion, while the integrity of anatomical structures is extremely important in medical images. For instance, errors in image translation may distort, shift, or even remove structures and tumors, leading to incorrect diagnosis and inadequate treatments. Training and conditioning diffusion models using paired source and target images with matching anatomy can help. However, such paired data are very difficult and costly to obtain, and may also reduce the robustness of the developed model to out-of-distribution testing data. We propose a frequency-guided diffusion model (FGDM) that employs frequency-domain filters to guide the diffusion model for structure-preserving image translation. Based on its design, FGDM allows zero-shot learning, as it can be trained solely on the data from the target domain, and used directly for source-to-target domain translation without any exposure to the source-domain data during training. We evaluated it on three cone-beam CT (CBCT)-to-CT translation tasks for different anatomical sites, and a cross-institutional MR imaging translation task. FGDM outperformed the state-of-the-art methods (GAN-based, VAE-based, and diffusion-based) in metrics of Fréchet Inception Distance (FID), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index Measure (SSIM), showing its significant advantages in zero-shot medical image translation.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/tmi.2023.3325703

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Figures

Image 1
CC-BY · From the PDF
Fig. 1 Examples of intensity domain and spatial frequency domain visualizations of three pairs of CBCT images and CT images from different datasets.
Fig. 1 Examples of intensity domain and spatial frequency domain visualizations of three pairs of CBCT images and CT images from different datasets.CC-BY · From the PDF
Fig. 2 This illustration presents the statistical data of the Mean Squared Error (MSE) of frequency amplitudes between paired CBCT and CT across various spatial frequencies, derived from the frequency domains of three distinct datasets, where the first row is the average MSE distribution for the entire dataset. To further demonstrate that the difference between paired CBCT and CT is mainly concentrated in the intermediate frequencies, we show in the second row the box plot of the distribution of the frequencies corresponding to the maximum peak of MSE for the whole dataset.
Fig. 2 This illustration presents the statistical data of the Mean Squared Error (MSE) of frequency amplitudes between paired CBCT and CT across various spatial frequencies, derived from the frequency domains of three distinct datasets, where the first row is the average MSE distribution for the entire dataset. To further demonstrate that the difference between paired CBCT and CT is mainly concentrated in the intermediate frequencies, we show in the second row the box plot of the distribution of the frequencies corresponding to the maximum peak of MSE for the whole dataset.CC-BY · From the PDF
Fig. 3 The CBCT-to-CT translation process via the frequency-guided diffusion model (FGDM). The low-frequency information and high-frequency information of CBCTs are used as the diffusion conditions, and the information of intermediate frequencies is gradually generated through the reverse diffusion process to yield translated target images in the CT domain.
Fig. 3 The CBCT-to-CT translation process via the frequency-guided diffusion model (FGDM). The low-frequency information and high-frequency information of CBCTs are used as the diffusion conditions, and the information of intermediate frequencies is gradually generated through the reverse diffusion process to yield translated target images in the CT domain.CC-BY · From the PDF
Fig. 4 The structure diagram of the frequency-guided diffusion model (FGDM). The model is trained solely on the target image domain (CT) to generate information of the intermediate frequency, conditioned on the low and high frequency CT information. For zero-shot CBCT-to-CT translation, the low and high frequency information of the CBCTs (source) is used instead as conditions to generate corresponding CT images. In the training phase, noise is added until the step T , while in the zero-shot translation stage, noise is applied until the step ˜ T ( ˜ T ≤ T ). Using a large T during the training stage is motivated by our desire to have the model adeptly handle a range of low-frequency information thresholds. Using a different ˜ T from T during the zero-shot translation stage allows us to customize the low-frequency information threshold according to the specific disparities between the source and target domains.
Fig. 4 The structure diagram of the frequency-guided diffusion model (FGDM). The model is trained solely on the target image domain (CT) to generate information of the intermediate frequency, conditioned on the low and high frequency CT information. For zero-shot CBCT-to-CT translation, the low and high frequency information of the CBCTs (source) is used instead as conditions to generate corresponding CT images. In the training phase, noise is added until the step T , while in the zero-shot translation stage, noise is applied until the step ˜ T ( ˜ T ≤ T ). Using a large T during the training stage is motivated by our desire to have the model adeptly handle a range of low-frequency information thresholds. Using a different ˜ T from T during the zero-shot translation stage allows us to customize the low-frequency information threshold according to the specific disparities between the source and target domains.CC-BY · From the PDF
Fig. 5 Visual comparison between GAN-based, VAE-based, and diffusion model-based CBCT-to-CT translation methods on the head and neck dataset (in-distribution). The display window is [-1000,1000] HU.
Fig. 5 Visual comparison between GAN-based, VAE-based, and diffusion model-based CBCT-to-CT translation methods on the head and neck dataset (in-distribution). The display window is [-1000,1000] HU.CC-BY · From the PDF
Fig. 6 Visual comparison between GAN-based, VAE-based, and diffusion model-based CBCT-to-CT translation methods on the lung dataset (out-of-distribution). The display window is [-1000,1000] HU.
Fig. 6 Visual comparison between GAN-based, VAE-based, and diffusion model-based CBCT-to-CT translation methods on the lung dataset (out-of-distribution). The display window is [-1000,1000] HU.CC-BY · From the PDF
Image 8
CC-BY · From the PDF
Image 9
CC-BY · From the PDF
Fig. 9 Visual comparison between images generated using different values of high-frequency ratio parameter η on the head and neck dataset. The display window is [-1000,1000] HU.
Fig. 9 Visual comparison between images generated using different values of high-frequency ratio parameter η on the head and neck dataset. The display window is [-1000,1000] HU.CC-BY · From the PDF
Fig. 10 Visualization comparison with different low frequency ratio (step ˜ T ) on the head and neck dataset. The display window is [-1000,1000] HU.
Fig. 10 Visualization comparison with different low frequency ratio (step ˜ T ) on the head and neck dataset. The display window is [-1000,1000] HU.CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 39 references

Research works printed but not linked: 1

  1. Song, Score-based generative modeling through stochastic differential equations, 1

Cited by

See all 128

Metadata

Authors
Li, Y., Shao, H.-C., Liang, X., Chen, L., Li, R., Jiang, S., Wang, J., & Zhang, Y.
Affiliations
The University of Texas Southwestern Medical Center (US)
Venue
IEEE Transactions on Medical Imaging
Published
2024-03
Volume
43
Issue
3
Pages
980-993
ISSN
  • 0278-0062 (ISSN-L)
  • 1558-254X
OpenAlex
W4387757480
Cited by
111
References
39
Citations by year
2026: 44 · 2025: 48 · 2024: 16
FWCI
10.07
Type
journal article
Language
en
Open access
Locations
  • IEEE Transactions on Medical Imaging (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (acceptedVersion)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
Funders
National Institutes of Health (R01 CA240808, R01 CA258987)
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
  • Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
diffusion models, frequency domain filters, structure preservation, out-of-distribution robustness
MeSH terms
Image Processing, Computer-Assisted, Magnetic Resonance Imaging, Phantoms, Imaging, Cone-Beam Computed Tomography, Signal-To-Noise Ratio

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Zero-Shot Medical Image Translation via Frequency-Guided Diffusion Models. https://app.cochat.ai/papers/10.1109/tmi.2023.3325703
The paper
Li, Y., Shao, H.-C., Liang, X., Chen, L., Li, R., Jiang, S., Wang, J., & Zhang, Y. (2024). Zero-Shot Medical Image Translation via Frequency-Guided Diffusion Models. IEEE Transactions on Medical Imaging. https://doi.org/10.1109/tmi.2023.3325703

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.