Sparse-view X-ray 3D reconstruction is essential for reducing radiation exposure, but recovering a density field from a handful of X-ray projections is severely ill-posed. Recently, 3D Gaussian Splatt...
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.
arXiv:2609.22849v1 Announce Type: new Abstract: In novel view synthesis and Computed Tomography (CT) reconstruction with sparse-view X-ray imaging, insufficient angular coverage leads to structural a...
FaCT-GS is a new framework that accelerates Gaussian Splatting (GS) for X‑ray Computed Tomography (CT) reconstruction. By optimizing voxelization and rasterization pipelines, it achieves speeds more than four times faster than the current state‑of‑the‑art GS methods on 512×512 projections and over thirteen times faster on 2k projections. The improved voxelization also allows quick fitting of Gaussians to existing volumes, providing a compressed representation or a warm‑start prior for reconstruction.
X‑Splat introduces a unified Gaussian splatting framework that jointly reconstructs RGB and X‑ray 3D scenes. By sharing a single set of geometric parameters across modalities and adding appearance attributes only to surface‑proximal primitives, the method avoids redundant geometry and alignment artifacts. Cross‑modal registration is achieved through clean multi‑view stereo and X‑ray geometric proxies, and a surface‑depth consistency loss guides RGB reconstruction, resulting in robust alignment, higher RGB rendering quality, and a more compact representation compared to separate models.
Sparse-view computed tomography is a severely ill-posed inverse problem, where recent 3D Gaussian Splatting methods offer an efficient explicit representation for tomographic reconstruction. However, we find that projection-domain optimization can be misleading in this setting: the rendered projections may continue to improve while the reconstructed volume deteriorates.
The paper introduces LiftXR, a geometry‑guided framework that first reconstructs a 3D anatomical layout from bi‑planar X‑ray images and then uses this layout to guide CT volume reconstruction. An anatomical parser refines the layout by analyzing the reconstructed CT, enabling region‑specific intensity calibration. Experiments on two public datasets show LiftXR surpasses recent X‑ray‑to‑CT methods and improves downstream segmentation performance.
The paper introduces a method for extremely sparse‑view computed tomography (xSV‑CT) that learns a pose‑aware diffusion prior from a single 2D projection per training sample. During inference, the diffusion process is guided by measured‑view fidelity, shared‑volume reprojection, and projection‑physics consistency to reconstruct 3D volumes from only 1, 4, or 8 views. Experiments on two public CT datasets show that this approach outperforms or matches baselines that rely on dense‑volume supervision, achieving superior novel‑view synthesis and competitive reconstruction quality.
arXiv:2607. 04661v1 Announce Type: cross Abstract: Reconstructing 3D scene structures from sparse, low-overlap observations remains a fundamental challenge in autonomous driving.
arXiv:2609.12343v1 Announce Type: new Abstract: Feed-forward Gaussian splatting models have demonstrated remarkable effectiveness in reconstructing three-dimensional (3D) objects from a few two-dimen...
The paper introduces $K$-NeAS, a scalable neural architecture for multi-material CT reconstruction that replaces separate material networks with a shared latent backbone and a differentiable $K$-material soft selector. It automates attenuation bounds using a Gaussian Mixture Model and adds a scheduled auxiliary floater loss to reduce geometric hallucinations in sparse-view settings. Evaluated on four clinical CBCT datasets, $K$-NeAS achieves higher 3D volumetric fidelity—up to a 1.88 dB PSNR gain over a single-material baseline—and shows improved robustness under extreme sparsity, outperforming baselines by up to 1.17 dB.
CompactSplat introduces a feedforward scene reconstruction method that adaptively allocates 3D Gaussian primitives based on texture-guided spatial partitioning and hierarchical sampling. By concentrating dense Gaussians in complex regions and using expanded primitives for flat areas, it reduces redundancy and better captures planar geometry. The framework also allows dynamic sparsity control during inference, enabling trade-offs between rendering quality and memory usage without retraining.