Metal Artifact Reduction
Metal artifact reduction (MAR) removes the dark and bright streaks that metal implants, dental fillings and surgical hardware cause in CT images. Metal attenuates X-rays so strongly that the rays passing through it suffer beam hardening, photon starvation, scatter and partial-volume errors, and the corrupted projections spread into streaks across the whole slice.
MAR matters in dental and maxillofacial CBCT, orthopaedic imaging around joint replacements and spinal fixation, and radiotherapy planning, where streaks distort the electron densities used for dose calculation.
Classical methods replace the metal trace in the sinogram by interpolation (linear interpolation, normalised MAR) or model the polychromatic physics iteratively. Deep-learning methods work in the image domain, the sinogram domain or both, often with interpretable model-driven designs. Unsupervised approaches, such as polychromatic neural representations and diffusion priors, remove the need for paired metal-free data, which is rarely available clinically.
Browse all 23 Metal Artifact Reduction papers in the searchable directory
Start here
If you are new to this topic, these are good first reads, chosen by the maintainers.
- Metal Artifact Reduction in CT: Where Are We After Four Decades? (IEEE Access, 2016)
A review of four decades of metal artifact reduction, from the physics to six classes of correction methods. - An Interpretable Dual Domain Network for CT Metal Artifact Reduction (MICCAI, 2021 · code)
An interpretable dual-domain network that unrolls a model-based MAR formulation, with code. - Adaptive Convolutional Dictionary Network for CT Metal Artifact Reduction (IJCAI, 2022 · code)
An adaptive convolutional dictionary network that keeps model-based interpretability in the image domain, with code. - Unsupervised Polychromatic Neural Representation for CT Metal Artifact Reduction (NeurIPS, 2023 · code)
An unsupervised polychromatic neural representation that models the metal physics directly and needs no paired data, with code. - Unsupervised CT Metal Artifact Reduction by Plugging Diffusion Priors in Dual Domains (TMI, 2024 · code)
Unsupervised MAR that plugs diffusion priors into both the image and the sinogram domain, with code.
Papers
All 23 papers, newest first.
Browse all 23 Metal Artifact Reduction papers in the searchable directory
Datasets, benchmarks and toolkits
Entries from the directory whose tasks or tags match this topic.
- CTPelvic1K (official page) · Dataset — 1,184 pelvic CT scans across 7 sub-datasets; notably includes metal-implant cases, one of the few open sources of real metal artefacts.
- Physics-ArX · Toolkit — Physics-based data-augmentation library that injects realistic CT/CBCT noise, scatter and artefact models into existing images to build paired training data.
- XrayPhysics · Toolkit — X-ray cross sections, source-spectrum modelling, multi-material beam-hardening correction and dual-energy decomposition. From the LEAP author, and it pairs naturally with a simulator when you need a physically honest spectrum.
Guides
Guides that cover this topic.
Other topics
- Surveys & Reviews (24)
- Classical & Iterative Reconstruction (71)
- Sparse-View CT (157)
- Limited-Angle Tomography (46)
- Low-Dose CT Denoising (69)
- Spectral & Dual-Energy CT (19)
- Cone-Beam CT (77)
- Dynamic & 4D CT (24)
- Scatter Correction (11)
- Geometry Calibration & Motion Compensation (28)
- Deep Unrolling & Model-Based Learning (68)
- Self-Supervised & Untrained Methods (35)
- Diffusion & Generative Priors (68)
- Implicit Neural Representations (25)
- Gaussian Splatting (26)