X-Splat: Gaussian Splatting for 3D CBCT Generation from Single Panoramic Radiograph
2026-07-02 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the challenge of creating detailed 3D dental images from a single 2D panoramic X-ray, which usually loses depth information. They introduce X-Splat, a new method that uses special Gaussian shapes placed along the known X-ray paths to rebuild the 3D structure more accurately. Their approach learns from simulated data and refines the 3D model using anatomy knowledge, outperforming previous methods by better capturing individual teeth and important bone features. They also provide new ways to evaluate these 3D reconstructions specifically for dental anatomy.
Panoramic Radiograph (PXR)Cone-Beam Computed Tomography (CBCT)Gaussian Splatting3D ReconstructionBeer-Lambert LawAnisotropic GaussiansVoxel-based VolumesNeural Radiance Fields (NeRF)Radiographic ImagingAnatomical Priors
Authors
Tomasz Szczepański, Szymon Płotka, Michal K. Grzeszczyk, Tomasz Trzciński, Arkadiusz Sitek
Abstract
Generating a 3D dental volume from a single panoramic radiograph (PXR) could provide a low-radiation alternative to Cone-Beam Computed Tomography (CBCT), but the problem is highly underdetermined: panoramic acquisition integrates 3D attenuation along curved X-ray paths into a 2D image, leaving depth-resolved anatomy unobserved. Existing implicit and generative approaches often produce oversmoothed geometry or anatomically inconsistent hallucinations, lacking geometry-driven supervision and relying on smooth representations unable to precisely localize sharp anatomical boundaries. We propose X-Splat, the first Gaussian Splatting framework for generating CBCT-like 3D dental volumes from a single PXR. X-Splat uses the known panoramic acquisition geometry as a generation scaffold: learnable anisotropic Gaussian primitives are initialized along the X-ray paths that formed the input image and adjusted in a single feed-forward pass, constrained by Beer-Lambert reprojection and multi-view radiographic training supervision. A lightweight residual refiner adds dataset-level anatomical priors without overriding the geometry already resolved by the Gaussians. We train on synthetic PXR-CBCT pairs, enabling direct volumetric supervision without paired real scans. We further introduce segmentation-based geometry-aware metrics, providing the first evaluation of PXR-based generation over maxillofacial anatomy. X-Splat outperforms NeRF- and GAN-based baselines, recovering individual teeth, cortical boundaries, and alveolar structure, including the mandibular canal which prior methods fail to reconstruct. Code will be available at https://github.com/tomek1911/X-Splat