EUVIMEDDescubrir EUVIMED
Crossref

Residual Metal Artifact Reduction in CT Images: An Unsupervised Residual and Contrastive Learning Approach for Preserving Metal Structures

YongSoo Kim, Jung-Woo Lee, Byung Chul Lee, Hyunseok Seo

Medical Physics · 2025

DOI: 10.1002/mp.70078 ↗

At a glance

Journal article in Medical Physics, 2025. Written by YongSoo Kim and 3 co-authors. Institution: Division of Bio-Medical Science & Technology, KIST School, Korea National University of Science and Technology, Seoul, South Korea. Major MEDLINE topics: Artifacts, Metals, Tomography, X-Ray Computed, Image Processing, Computer-Assisted. Cited 1 times so far (Crossref); 45 references. Funded by: KIST Institutional Programs, Ministry of Science and ICT, South Korea (MSIT).

ResumenSubjectsDatos bibliográficosCiteRelated studies

Abstract (original language)

Background

It is easy to find computed tomography (CT) images that include metals such as implants, bone plates, and bone shafts. These metals replacing body parts, cause serious artifacts in the CT image originated by x‐ray beam‐hardening. Traditionally, CT physics‐based image processing techniques were empirically applied to reduce metal artifacts (MAs). Recently, with the great success of deep learning, many studies on metal artifact reduction (MAR) using convolutional neural networks (CNNs) have been introduced.

Purpose

Most of them commonly meet a challenge to obtain ground truth images for MAR in clinical practice. Thus, for effective MAR without ground truth images, we propose a residual MA model for unsupervised deep learning scheme in CT images.

Methods

In the 1st stage, MAs are extracted by CT physics‐inspired residual model, which is enabled by the residual learning scheme. The result of the 1st stage is fed into the input for the 2nd stage, where artifacts that are not reduced enough in the 1st stage are further removed by the contrastive learning scheme. Then, the networks in the 2nd stage can easily recognize the original structures due to primary artifact reduction in the 1st stage and can properly refine the image.

Results

Our model was validated on three datasets. The results show that the proposed model outperforms other MAR models, preserving both original body and metal structures while reducing MAs effectively.

Conclusions

We hope that this unsupervised learning model can contribute to good achievements in the MAR field while overcoming the limitations of data construction.

Subjects

MeSH terms (MEDLINE)

ArtifactsMetalsTomography, X-Ray ComputedImage Processing, Computer-AssistedUnsupervised Machine LearningDeep LearningHumansNeural Networks, Computer

Author keywords

CT Physicsmetal artifact reductionunsupervised learning

Substances

Metals

Authors and institutions

  • Division of Bio-Medical Science & Technology, KIST School, Korea National University of Science and Technology, Seoul, South Korea.
  • Bionics Research Center, Biomedical Research Division, Korea Institute of Science and Technology, Seoul, South Korea.
  • Department of Oral and Maxillofacial Surgery, School of Dentistry, Kyung Hee University, Seoul, South Korea.
  • KHU-KIST Department of Converging Science and Technology, Graduate School, Kyung Hee University, Seoul, South Korea.
  • Department of Artificial Intelligence, Korea University, Seoul, South Korea.

Funding

  • KIST Institutional Programs
  • Ministry of Science and ICT, South Korea (MSIT)

Cite

APA

Kim, Y., Lee, J. W., Lee, B. C., & Seo, H. (2025). Residual Metal Artifact Reduction in CT Images: An Unsupervised Residual and Contrastive Learning Approach for Preserving Metal Structures. Medical Physics. https://doi.org/10.1002/mp.70078

Vancouver

Kim Y, Lee JW, Lee BC, Seo H. Residual Metal Artifact Reduction in CT Images: An Unsupervised Residual and Contrastive Learning Approach for Preserving Metal Structures. Medical Physics. 2025. doi:10.1002/mp.70078

BibTeX
@article{Kim2025,
  author = {Kim, YongSoo and Lee, Jung-Woo and Lee, Byung Chul and Seo, Hyunseok},
  title = {Residual Metal Artifact Reduction in CT Images: An Unsupervised Residual and Contrastive Learning Approach for Preserving Metal Structures},
  journal = {Medical Physics},
  year = {2025},
  doi = {10.1002/mp.70078}
}
Start for free →