Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

3D CBCT Artefact Removal Using Perpendicular Score-Based Diffusion Models

Susanne Schaub, Florentin Bieder, Matheus L. Oliveira, Yulan Wang, Dorothea Dagassan-Berndt, Michael M. Bornstein, Philippe C. Cattin

Abstract

Cone-beam computed tomography (CBCT) is a widely used 3D imaging technique in dentistry, offering high-resolution images while minimising radiation exposure for patients. However, CBCT is highly susceptible to artefacts arising from high-density objects such as dental implants, which can compromise image quality and diagnostic accuracy. To reduce artefacts, implant inpainting in the sequence of projections plays a crucial role in many artefact reduction approaches. Recently, diffusion models have achieved state-of-the-art results in image generation and have widely been applied to image inpainting tasks. However, to our knowledge, existing diffusion-based methods for implant inpainting operate on independent 2D projections. This approach neglects the correlations among individual projections, resulting in inconsistencies in the reconstructed images. To address this, we propose a 3D dental implant inpainting approach based on perpendicular score-based diffusion models, each trained in two different planes and operating in the projection domain. The 3D distribution of the projection series is modelled by combining the two 2D score-based diffusion models in the sampling scheme. Our results demonstrate the method's effectiveness in producing high-quality, artefact-reduced 3D CBCT images, making it a promising solution for improving clinical imaging.

3D CBCT Artefact Removal Using Perpendicular Score-Based Diffusion Models

Abstract

Cone-beam computed tomography (CBCT) is a widely used 3D imaging technique in dentistry, offering high-resolution images while minimising radiation exposure for patients. However, CBCT is highly susceptible to artefacts arising from high-density objects such as dental implants, which can compromise image quality and diagnostic accuracy. To reduce artefacts, implant inpainting in the sequence of projections plays a crucial role in many artefact reduction approaches. Recently, diffusion models have achieved state-of-the-art results in image generation and have widely been applied to image inpainting tasks. However, to our knowledge, existing diffusion-based methods for implant inpainting operate on independent 2D projections. This approach neglects the correlations among individual projections, resulting in inconsistencies in the reconstructed images. To address this, we propose a 3D dental implant inpainting approach based on perpendicular score-based diffusion models, each trained in two different planes and operating in the projection domain. The 3D distribution of the projection series is modelled by combining the two 2D score-based diffusion models in the sampling scheme. Our results demonstrate the method's effectiveness in producing high-quality, artefact-reduced 3D CBCT images, making it a promising solution for improving clinical imaging.
Paper Structure (15 sections, 7 equations, 3 figures, 1 table, 1 algorithm)

This paper contains 15 sections, 7 equations, 3 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background
  3. Score-Based Diffusion Models
  4. Diffusion Posterior Sampling
  5. Method
  6. Experiments
  7. Experimental Settings
  8. Dataset
  9. Artefact Simulation
  10. Implementation Details
  11. Results
  12. Comparing Methods
  13. Evaluation
  14. Discussion and Conclusion

Figures (3)

  • Figure 1: A primary diffusion model is trained on projections along the original projection direction, while a secondary model is trained on perpendicular projections. During sampling, the algorithm alternates between the two models to generate a coherent 3D image.
  • Figure 2: We computed the projection series from our implant-free CBCT reconstructions. Implants were added with a mask, which serves as input to the TPDM.
  • Figure 3: Top: CBCT reconstructions of an exemplary image. Bottom: Difference maps with respect to the reference image. The images used for the difference maps are in the range [-0.04, 0.07].