Learning Wavelet-Sparse FDK for 3D Cone-Beam CT Reconstruction

2025-05-19

AI Summary

To address the high noise sensitivity, severe artifacts, and trade-off between interpretability and efficiency in low-dose cone-beam CT (CBCT) reconstruction using the Feldkamp-Davis-Kress (FDK) algorithm, this paper proposes an interpretable lightweight neural network architecture. Preserving the classical FDK workflow structure, our method embeds learnable modules into the cosine-weighting and filtering stages and—novelty—the first to incorporate wavelet sparsification into learnable parameter modeling. This design reduces model parameters by 93.75%, maintains inference speed identical to standard FDK, and enables plug-and-play clinical deployment. Experimental results demonstrate substantial improvements in reconstructed image signal-to-noise ratio and noise robustness, while fully retaining the physical interpretability and computational efficiency inherent to FDK.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Calibration & Uncertainty Quantification
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Cone-Beam Computed Tomography (CBCT) is essential in medical imaging, and the Feldkamp-Davis-Kress (FDK) algorithm is a popular choice for reconstruction due to its efficiency. However, FDK is susceptible to noise and artifacts. While recent deep learning methods offer improved image quality, they often increase computational complexity and lack the interpretability of traditional methods. In this paper, we introduce an enhanced FDK-based neural network that maintains the classical algorithm's interpretability by selectively integrating trainable elements into the cosine weighting and filtering stages. Recognizing the challenge of a large parameter space inherent in 3D CBCT data, we leverage wavelet transformations to create sparse representations of the cosine weights and filters. This strategic sparsification reduces the parameter count by $93.75%$ without compromising performance, accelerates convergence, and importantly, maintains the inference computational cost equivalent to the classical FDK algorithm. Our method not only ensures volumetric consistency and boosts robustness to noise, but is also designed for straightforward integration into existing CT reconstruction pipelines. This presents a pragmatic enhancement that can benefit clinical applications, particularly in environments with computational limitations.

Problem

• Enhance FDK algorithm for CBCT to reduce noise and artifacts
• Maintain interpretability while integrating trainable neural elements
• Reduce parameter space using wavelet sparsity without performance loss

Innovation

• Enhanced FDK with trainable cosine weighting and filtering
• Wavelet sparsification reduces parameters by 93.75%
• Maintains FDK's interpretability and computational efficiency