1887

Abstract

Summary

This study investigates the effectiveness of different U-Net variants for suppressing ring artefacts in high-resolution X-ray microtomography (XMT) images of basalt from triaxial rock deformation experiment. Ring artefacts, due to electronic noise in detector diodes, distort greyvalue distributions, impair feature segmentation and digital volume correlation (DVC) accuracy. To overcome limited availability of XMT data, a custom Ring Artefact Generator (RAG) was developed to simulate artefacts of varying frequency and amplitude in sinograms, producing over 130,000 pairs of noisy and clean overlapping image patches for training. Three machine learning models: (a) U-Net (baseUNET), (b) Residual U-Net (ResUNET), and (c) Attention-Guided ResUNET (AG-ResUNET)—were trained and evaluated using PSNR and SSIM metrics.

Results show that all models effectively suppressed ring artefacts, with ResUNET achieving the highest PSNR improvement (57.5%) with comparable SSIM scores (0.99) across all models. The inclusion of residual connections accelerated convergence and improved denoising compared to the baseUNET, while attention mechanisms increased computational cost without significant performance gains due to limited data. The findings highlight that residual learning offers the best balance between denoising performance, model complexity, and efficiency, without impacting XMT images for downstream tasks like pore segmentation and DVC.

© EAGE Publications BV
Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.202639084
2026-03-09
2026-02-14

  • From This Site

    /content/papers/10.3997/2214-4609.202639084
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Journal -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

References

  1. Fu, T., Wang, Y., Zhang, K., Zhang, J., Wang, S., Huang, W., Wang, Y., Yao, C., Zhou, C., & Qin, Y. [2023]. Deep-learning-based ring artifact correction for tomographic reconstruction. Journal of Synchrotron Radiation, 30(3).
     [Google Scholar]
  2. He, K., Zhang, X., Ren, S., & Sun, J. [2016]. Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770–778.
     [Google Scholar]
  3. Jetley, S., Lord, N.A., Lee, N., & Torr, P.H.S. [2018]. Learn to Pay Attention. International Conference on Learning Representations (ICLR).
     [Google Scholar]
  4. Li, Y., Han, S., Zhao, Y., Li, F., Ji, D., Zhao, X., Liu, D., Jian, J., & Hu, C. [2022]. Synchrotron microtomography image restoration via regularization representation and deep CNN prior. Computer Methods and Programs in Biomedicine, 226, 107181.
     [Google Scholar]
  5. Mahdaviara, M., Mousavi, M., Rafiei, Y., Raoof, A., & Sharifi, M. [2025]. Improving numerical fluid flow simulation by ring artifact removal in micro-CT images of porous media using attention autoencoder–decoders. Transport in Porous Media, 152, 57.
     [Google Scholar]
  6. Oktay, O., et al. [2018]. Attention U-Net: Learning Where to Look for the Pancreas. Medical Image Computing and Computer-Assisted Intervention (MICCAI), 311–319.
     [Google Scholar]
  7. Paraskevoulakos, C., Ghosh, S., Andriollo, T., & Michel, A. [2024]. Sensitivity Study Using Synthetic 3D Image Datasets to Investigate the Effect of Noise Artefacts on Digital Volume Correlation. Experimental Mechanics, 64, 595–624.
     [Google Scholar]
  8. Ronneberger, O., Fischer, P., & Brox, T. [2015]. U- Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention (MICCAI), 234–241.
     [Google Scholar]
  9. Roy, A.G., Siddiqui, S., Pölsterl, S., Navab, N., & Wachinger, C. [2020]. Brain Tumor Segmentation Using Attention-Guided Residual U-Net. MICCAI BraTS Challenge.
     [Google Scholar]
  10. Schlüter, S., Sheppard, A., Brown, K., & Wildenschild, D. [2014]. Image processing of multiphase images obtained via X-ray microtomography: A review. Water Resources Research, 50, 3615–3639.
     [Google Scholar]
  11. Wang, Q., et al. [2020]. ECA- Net: Efficient Channel Attention for Deep Convolutional Neural Networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 11531–11539.
     [Google Scholar]
  12. Wang, Z., & Bovik, A.C. [2009]. Mean Squared Error: Love It or Leave It? A New Look at Signal Fidelity Measures. IEEE Signal Processing Magazine, 26(1), 98–117.
     [Google Scholar]
  13. Zhang, K., Zuo, W., Chen, Y., Meng, D., & Zhang, L. [2017]. Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising. IEEE Transactions on Image Processing, 26(7), 3142–3155.
     [Google Scholar]
/content/papers/10.3997/2214-4609.202639084
Loading
Evaluating UNet Variants for Efficient Ring Artefact Suppression in High-Resolution XMT Basalt Imaging
Conference Proceedings 2026, 1 (2026); https://doi.org/10.3997/2214-4609.202639084
/content/papers/10.3997/2214-4609.202639084
/content/papers/10.3997/2214-4609.202639084
Loading

Data & Media loading...

Most Cited This Month Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error