Beam Geometry and Input Dimensionality: Impact on Sparse-Sampling Artifact Correction for Clinical CT with U-Nets

2025-08-25

AI Summary

This study systematically investigates how beam geometry (parallel-, fan-, and cone-beam) and input dimensionality (2D, 2.5D, 3D) affect the performance of deep learning–based sparse-sampling streak artifact correction in clinical CT. We propose a 2.5D input representation that incorporates local 3D contextual information via multi-planar slice recombination and design a corresponding U-Net architecture, evaluating all configurations uniformly on Astra Toolbox–simulated data. Contrary to expectations, the axial 2D U-Net—trained solely on 2D slices—achieves significantly lower MSE and higher SSIM than both 2.5D and 3D counterparts, while also demonstrating superior generalization. This challenges the common assumption that higher-dimensional inputs inherently improve reconstruction fidelity, revealing instead that local texture modeling is more critical than global geometric modeling for artifact correction. The findings provide empirical evidence supporting lightweight, computationally efficient CT denoising architectures optimized for clinical deployment.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Stochastic Optimization
• Computer Vision: Representation Learning for Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

This study aims to investigate the effect of various beam geometries and dimensions of input data on the sparse-sampling streak artifact correction task with U-Nets for clinical CT scans as a means of incorporating the volumetric context into artifact reduction tasks to improve model performance. A total of 22 subjects were retrospectively selected (01.2016-12.2018) from the Technical University of Munich's research hospital, TUM Klinikum rechts der Isar. Sparsely-sampled CT volumes were simulated with the Astra toolbox for parallel, fan, and cone beam geometries. 2048 views were taken as full-view scans. 2D and 3D U-Nets were trained and validated on 14, and tested on 8 subjects, respectively. For the dimensionality study, in addition to the 512x512 2D CT images, the CT scans were further pre-processed to generate a so-called '2.5D', and 3D data: Each CT volume was divided into 64x64x64 voxel blocks. The 3D data refers to individual 64-voxel blocks. An axial, coronal, and sagittal cut through the center of each block resulted in three 64x64 2D patches that were rearranged as a single 64x64x3 image, proposed as 2.5D data. Model performance was assessed with the mean squared error (MSE) and structural similarity index measure (SSIM). For all geometries, the 2D U-Net trained on axial 2D slices results in the best MSE and SSIM values, outperforming the 2.5D and 3D input data dimensions.

Problem

• Investigating beam geometry effects on CT artifact correction
• Assessing input data dimensionality impact on U-Net performance
• Improving sparse-sampling streak artifact reduction in clinical CT

Innovation

• U-Nets for CT artifact correction
• Beam geometry impact on model performance
• 2D U-Net outperforms 2.5D and 3D inputs