Data-Efficient Limited-Angle CT Using Deep Priors and Regularization

2025-02-17

AI Summary

In limited-angle CT (e.g., with only a few projection views), the inverse Radon problem is severely ill-posed, leading to pronounced reconstruction artifacts. Method: We propose an ultra-low-data paradigm requiring merely 12 real measurement points—8 for prior learning and 4 for hyperparameter tuning—without any large-scale labeled training data. Our approach integrates a differentiable Radon transform, total variation regularization, sinogram-domain filtering, deep image prior (DIP), and a patch-level autoencoder, augmented by multi-scale regularization to enhance structural fidelity. Results: Evaluated on the Helsinki Tomography Challenge 2022 dataset, our method achieves reconstruction quality comparable to state-of-the-art synthetic-data-driven approaches, significantly improving image fidelity and generalizability under extreme undersampling. This establishes a novel paradigm for clinical CT imaging in radiation-sensitive scenarios.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Reconstructing an image from its Radon transform is a fundamental computed tomography (CT) task arising in applications such as X-ray scans. In many practical scenarios, a full 180-degree scan is not feasible, or there is a desire to reduce radiation exposure. In these limited-angle settings, the problem becomes ill-posed, and methods designed for full-view data often leave significant artifacts. We propose a very low-data approach to reconstruct the original image from its Radon transform under severe angle limitations. Because the inverse problem is ill-posed, we combine multiple regularization methods, including Total Variation, a sinogram filter, Deep Image Prior, and a patch-level autoencoder. We use a differentiable implementation of the Radon transform, which allows us to use gradient-based techniques to solve the inverse problem. Our method is evaluated on a dataset from the Helsinki Tomography Challenge 2022, where the goal is to reconstruct a binary disk from its limited-angle sinogram. We only use a total of 12 data points--eight for learning a prior and four for hyperparameter selection--and achieve results comparable to the best synthetic data-driven approaches.

Problem

• Limited-angle CT reconstruction
• Reduced radiation exposure
• Deep priors and regularization methods

Innovation

• Combines Total Variation regularization
• Uses Deep Image Prior
• Implements differentiable Radon transform