HiGDiff is proposed, a feed-forward hierarchical Gaussian diffusion framework that decomposes reconstruction both spatially and from structure to detail in three distinct CT benchmark datasets.
Abstract
Reconstructing three-dimensional computed tomography (CT) from severely constrained projections is highly ill-posed. Sparse angular sampling, restricted angular coverage, and low photon counts can occur individually or jointly, obscuring global anatomy and local tissue detail. Many learned CT reconstruction methods are tailored to a single dominant degradation. Existing diffusion and Gaussian approaches commonly recover global structure and local detail within a shared representation. We propose HiGDiff, a feed-forward hierarchical Gaussian diffusion framework that decomposes reconstruction both spatially and from structure to detail. Physics-conditioned anatomical anchors and a foreground capacity field allocate learnable Gaussian primitives to informative regions. A structure diffusion stage first recovers global attenuation geometry, and its learned representation conditions a detail diffusion stage for residual boundaries and tissue transitions. The resulting Gaussian banks are rendered as attenuation fields and further refined by a gradient-isolated residual module. Experiments on three distinct CT benchmark datasets demonstrate state-of-the-art reconstruction performance across isolated, paired, and joint degradation settings, including improvements of 5.81 dB in macro-average peak signal-to-noise ratio (PSNR) and 0.113 in structural similarity index measure (SSIM) on the Low Dose CT Image and Projection Data (LDCT-PD) collection. Code and experimental configurations are openly available at https://github.com/Bean-Young/HiGDiff.
This paper proposes K-NeAS, a unified and scalable architecture for automated, multi-material surface reconstruction that replaces independent material networks with a shared latent backbone and introduces a fully differentiable $K$-material sequential soft selector to model an arbitrary number of overlapping tissues.
Daksh K. Shah, Emmanouil Nikolakakis, Razvan V. Marinescu· 0 citations
Sparse-view computed tomography (CT) reconstruction aims to recover high-quality CT volumes from a limited number of X-ray projection images, thereby reducing radiation exposure during image acquisition. However, this problem is inherently ill-posed because each projection provides only indirect line-integral supervision, and different attenuation distributions can explain similar sparse measurements. Existing analytic and iterative methods often suffer from streak artifacts and unstable solutions, while supervised learning-based methods require paired training data and may generalize poorly across anatomical regions or acquisition settings. Neural Radiance Field (NeRF)-based methods have recently shown promise by representing the attenuation field as a continuous coordinate-based function optimized directly from projection images. Nevertheless, these methods mainly enforce projection consistency and do not explicitly use volume-domain uncertainty to guide subsequent reconstruction. In this work, we propose EpiC-NeRF, a CT-specific closed-loop framework that actively feeds estimated epistemic uncertainty back into sparse-view reconstruction. EpiC-NeRF adapts evidential uncertainty estimation and aggregation to the X-ray CT line-integral formulation and maintains the resulting spatial uncertainty in a persistent three-dimensional Epistemic Grid Map. The accumulated uncertainty is used by Epistemic-Adaptive Layer Normalization to modulate intermediate features and by dual active sampling to guide ray- and point-level sample allocation. The newly estimated uncertainty then updates the grid map and guides subsequent optimization iterations, forming a unified feedback loop between uncertainty estimation and CT reconstruction. Experiments on four CT volume datasets demonstrate that EpiC-NeRF achieves improved reconstruction fidelity over existing analytic, iterative, and neural implicit reconstruction methods.
Donghyuk Choo, Haill An, Younhyun Jung· Mathematics· 0 citations
Sparse-view computed tomography (CT) reduces radiation dose by acquiring fewer projection views, but the resulting inverse problem is highly ill-posed and often produces severe streak artifacts. Existing deep reconstruction methods have achieved promising performance, yet many rely on first-order updates or large regularization networks, which can be less effective in ill-conditioned settings. We propose \textbf{CG-GLORE}, a compact deep unrolling framework inspired by second-order optimization for sparse-view CT reconstruction. Each unrolled stage uses a CG-solved linear system based on a structured Hessian surrogate: it retains the physics-induced curvature of the data-fidelity term while using an identity approximation for the learned regularization term. Thus, the method is second-order-inspired rather than an exact Newton method for the full learned objective. To model image priors, we design a Global-Local Regularization Network (GLORE), which combines convolutional local feature extraction with a Long-Range Dependency Representation module based on sparse patchification and Nystr\"{o}m attention. This design captures anatomical details and non-local dependencies while maintaining practical complexity. Experiments on AAPM and DeepLesion under multiple sparse-view and noise settings show that CG-GLORE achieves strong quantitative performance, stable convergence, lower noise power, and improved visual fidelity compared with representative reconstruction methods.
Tran Xuan Hieu Le, D. C. Bui, V. Le et al.· 0 citations
Computed tomography (CT) throughput is limited by scan time, which grows with both the number of projections acquired and the detector integration time for each. Reconstructing high-quality volumes from sparse-view or low-dose measurements therefore depends on an informative prior, typically a neural network trained for one specific scan setting and retrained whenever the modality, geometry, or material changes. We investigate whether a single diffusion model trained across several imaging domains can instead serve as a prior for many CT problems simultaneously. We evaluate the proposed method using the same frozen model on three datasets that differ in modality, beam geometry, material, and degradation type, spanning flaw analysis in additively manufactured metal parts imaged with cone-beam X-ray CT and concrete microstructure imaged with parallel-beam neutron CT. Our proposed method out-performs analytic reconstructions in all three cases, providing a step toward a reusable foundation prior for heterogeneous CT reconstruction problems.
Haley Duba-Sullivan, Patxi Fernandez-Zelaia, Obaidullah Rahman et al.· 0 citations
Computed tomography (CT) reconstruction under sparse-view acquisition is fundamentally ill-posed. Recently, 3D Gaussian Splatting (3DGS) has emerged as an efficient alternative to implicit neural fields for per-case tomographic reconstruction, offering explicit geometric primitives and fast differentiable rendering. However, under sparse projection supervision, existing 3DGS-based CT methods can become unstable because the Gaussian primitives are optimized largely independently, often leading to overfitting and needle-like artifacts. In this work, we propose GR-Gaussian, a graph-regularized radiative 3DGS framework for sparse-view CT reconstruction. Rather than introducing a new volumetric representation, our method augments radiative Gaussian optimization with local graph-guided structural cues. Specifically, we introduce two components: (1) a denoised point-cloud initialization strategy (De-Init), which filters artifact-contaminated FDK priors to provide a more reliable initialization for Gaussian placement and neighborhood construction; and (2) a Pixel-Graph-Aware (PGA) densification criterion, which supplements the baseline pixel-aware densification signal with local density contrast measured on the Gaussian neighborhood graph. In addition, we incorporate graph Laplacian and volumetric total variation regularization to improve structural consistency during optimization. Experiments on the X-3D and real-world CT datasets show that GR-Gaussian consistently improves reconstruction quality over the evaluated baselines, while providing cleaner structures and stronger suppression of sparse-view artifacts. These results indicate that initialization and graph-guided densification are effective practical extensions to radiative 3DGS for sparse-view CT reconstruction.
Yikuang Yuluo, Kuan Shen, Yue Ma et al.· IEEE Transactions on Computa...· 0 citations
X-ray imaging can be approximately modeled as the projection of an underlying volumetric attenuation field, with each measurement recording the accumulated attenuation along a corresponding ray path. Reconstructing a CT volume from only a few X-ray views is therefore severely ill-posed, as the projections collapse depth information and leave 3D locations of anatomical regions and their corresponding intensity distributions highly entangled and ambiguous. We observe that once the spatial organization of anatomical regions is established, estimating their CT intensities becomes substantially more tractable. Motivated by this, we propose LiftXR, an interleaved, geometry-guided framework that explicitly incorporates spatial layout recovery into CT reconstruction. Specifically, a layout lifter first generates a 3D anatomical layout from bi-planar X-rays, providing spatial guidance for an intensity renderer to reconstruct a CT volume. An anatomical parser then performs volumetric perception on the reconstruction, exploiting its spatially resolved boundary and intensity cues to recover a refined anatomical layout. This transition from projection-conditioned layout generation to reconstruction-conditioned anatomical perception allows the parsed layout to provide feedback for region-specific intensity calibration. Extensive experiments on two public datasets demonstrate that LiftXR consistently outperforms recent X-ray-to-CT reconstruction methods, establishing a new state of the art. Moreover, the reconstructed CT achieves superior performance in external downstream segmentation, indicating improved anatomical fidelity. Code will be released.
Yifei Wu, Yicheng Wu, Qiang Ma et al.· 0 citations