Abstract:Sparse-view Computed Tomography (CT) is an emerging protocol designed to reduce X-ray dose radiation in medical imaging. Traditional Filtered Back Projection algorithm reconstructions suffer from severe artifacts due to sparse data. In contrast, Model-Based Iterative Reconstruction (MBIR) algorithms, though better at mitigating noise through regularization, are too computationally costly for clinical use. This paper introduces a novel technique, denoted as the Deep Guess acceleration scheme, using a trained neural network both to quicken the regularized MBIR and to enhance the reconstruction accuracy. We integrate state-of-the-art deep learning tools to initialize a clever starting guess for a proximal algorithm solving a non-convex model and thus computing an interpretable solution image in a few iterations. Experimental results on real CT images demonstrate the Deep Guess effectiveness in (very) sparse tomographic protocols, where it overcomes its mere variational counterpart and many data-driven approaches at the state of the art. We also consider a ground truth-free implementation and test the robustness of the proposed framework to noise.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of image reconstruction in Sparse - View CT (abbreviated as sparse - view CT). Specifically, the author attempts to improve the speed and accuracy of sparse - view CT image reconstruction by introducing a new technique - the Deep Guess acceleration scheme. #### Background problems of sparse - view CT 1. **The need for radiation dose reduction**: - Traditional CT imaging requires a large amount of X - ray projection data to obtain high - quality images. However, in order to reduce the radiation dose received by patients, researchers hope to use fewer projection angles (i.e., sparse - view angles) for imaging. 2. **Limitations of traditional reconstruction algorithms**: - **Filtered Back Projection (FBP)**: This is a classic analytical algorithm, but it will introduce severe artifacts (such as streak artifacts) in the sparse - view case, resulting in a decline in image quality. - **Model - Based Iterative Reconstruction (MBIR)**: This method can better suppress noise through mathematical models and regularization techniques, but the computational cost is very high and it is difficult to be widely applied in clinical practice. #### Solutions proposed in the paper To solve the above problems, the paper proposes the **Deep Guess acceleration scheme**, whose core idea is to combine the advantages of deep learning and MBIR: 1. **Using deep learning for initial guess**: - Use a trained neural network to perform preliminary reconstruction of CT images in the sparse - view case, generating an initial guess (called Deep Guess) close to the real solution. This initial guess can significantly reduce the number of steps required for subsequent iterative optimization. 2. **Accelerating the solution of non - convex optimization problems**: - Take Deep Guess as the initial value and apply it to solve the optimization problem of non - convex T pV (Total p - norm Variation) regularization. In this way, high - quality reconstruction results can be obtained within a fewer number of iterations and avoid falling into bad local minima. 3. **Maintaining mathematical interpretability**: - Although deep learning is used, the final reconstruction result is still obtained through the MBIR algorithm, so the mathematical interpretability and convergence properties are retained. #### Experimental verification The paper verifies the effectiveness of the Deep Guess acceleration scheme through experiments. Especially under very sparse - view protocols, this method can outperform existing variational methods and other data - driven methods. In addition, the implementation method without the need for true labels is also considered, and the robustness of the framework to noise is tested. In summary, the main goal of this paper is to develop a fast, reliable, and interpretable sparse - view CT image reconstruction tool to meet the needs for efficient and high - quality image reconstruction in clinical applications.

Deep Guess acceleration for explainable image reconstruction in sparse-view CT

Deep Fusion Network Based Sparse View CT Reconstructions for Clinical Diagnostic Scanners

Generalized Deep Iterative Reconstruction for Sparse-View CT Imaging

Combining Deep Learning and Adaptive Sparse Modeling for Low-dose CT Reconstruction

An Educated Warm Start For Deep Image Prior-Based Micro CT Reconstruction

LRR-CED: low-resolution reconstruction-aware convolutional encoder-decoder network for direct sparse-view CT image reconstruction

Deep Convolutional Dictionary Learning Network for Sparse View CT Reconstruction with a Group Sparse Prior

Enhancing Low-dose CT Image Reconstruction by Integrating Supervised and Unsupervised Learning

An Efficient Sinogram Domain Fully Convolutional Interpolation Network for Sparse-View Computed Tomography Reconstruction

A model-guided deep network for limited-angle computed tomography

LIR-Net:Learnable Iterative Reconstruction Network for Fan Beam CT Sparse-View Reconstruction

Computationally efficient deep neural network for computed tomography image reconstruction

Sparse-view and limited-angle CT reconstruction with untrained networks and deep image prior

Unsupervised Sparse-view Backprojection via Convolutional and Spatial Transformer Networks

A Deep Learning Reconstruction Method for Sparsely Spaced Multidetector Helical CT Using Cascaded Dual-Domain CNN

Deep-neural-network based sinogram synthesis for sparse-view CT image reconstruction

Iterative reconstruction for limited-angle CT using implicit neural representation

DREAM-Net: Deep Residual Error Iterative Minimization Network for Sparse-View CT Reconstruction

Accurate and robust sparse-view angle CT image reconstruction using deep learning and prior image constrained compressed sensing (DL-PICCS)

Implicit Neural Representations for Robust Joint Sparse-View CT Reconstruction

Linear diffusion noise boosted deep image prior for unsupervised sparse-view CT reconstruction