Ultrafast Deep Learning-Based Scatter Estimation in Cone-Beam Computed Tomography

2025-09-10

AI Summary

Scatter artifacts severely degrade image quality in cone-beam computed tomography (CBCT), yet existing deep learning–based scatter estimation methods are computationally prohibitive for edge devices such as mobile CBCT systems due to excessive model size. To address this, we propose a lightweight multi-resolution deep network that innovatively identifies and exploits the critical role of downsampling in scatter modeling. By employing bilinear interpolation to construct multi-scale inputs, our approach significantly reduces computational and memory overhead while improving estimation accuracy. Experiments demonstrate that, compared to baseline methods, our model achieves a 78× reduction in FLOPs, a 16× speedup in inference time, and a 12× decrease in GPU memory usage, while attaining a mean absolute percentage error (MAPE) of 3.85% and a mean squared error (MSE) of 1.34×10⁻². The method exhibits robust performance on both synthetic and real-world CBCT data, enabling, for the first time, high-accuracy scatter estimation on resource-constrained edge platforms.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Medical and Biological Imaging
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Purpose: Scatter artifacts drastically degrade the image quality of cone-beam computed tomography (CBCT) scans. Although deep learning-based methods show promise in estimating scatter from CBCT measurements, their deployment in mobile CBCT systems or edge devices is still limited due to the large memory footprint of the networks. This study addresses the issue by applying networks at varying resolutions and suggesting an optimal one, based on speed and accuracy. Methods: First, the reconstruction error in down-up sampling of CBCT scatter signal was examined at six resolutions by comparing four interpolation methods. Next, a recent state-of-the-art method was trained across five image resolutions and evaluated for the reductions in floating-point operations (FLOPs), inference times, and GPU memory requirements. Results: Reducing the input size and network parameters achieved a 78-fold reduction in FLOPs compared to the baseline method, while maintaining comarable performance in terms of mean-absolute-percentage-error (MAPE) and mean-square-error (MSE). Specifically, the MAPE decreased to 3.85% compared to 4.42%, and the MSE decreased to 1.34 imes 10^{-2} compared to 2.01 imes 10^{-2}. Inference time and GPU memory usage were reduced by factors of 16 and 12, respectively. Further experiments comparing scatter-corrected reconstructions on a large, simulated dataset and real CBCT scans from water and Sedentex CT phantoms clearly demonstrated the robustness of our method. Conclusion: This study highlights the underappreciated role of downsampling in deep learning-based scatter estimation. The substantial reduction in FLOPs and GPU memory requirements achieved by our method enables scatter correction in resource-constrained environments, such as mobile CBCT and edge devices.

Problem

• Reducing deep learning network memory for CBCT scatter estimation
• Optimizing resolution balance between speed and accuracy
• Enabling scatter correction on mobile and edge devices

Innovation

• Multi-resolution deep learning networks
• Downsampling reduces computational requirements
• Optimized for mobile and edge devices