Limited-angle tomographic reconstruction is one of the typical ill-posed inverse problems, leading to edge divergence with degraded image quality. Recently, deep learning has been introduced into image reconstruction and achieved great results. However, existing deep reconstruction methods have not fully explored data consistency, resulting in poor performance. In addition, deep reconstruction methods are still mathematically inexplicable and unstable. In this work, we propose an iterative residual optimization network (IRON) for limited-angle tomographic reconstruction. First, a new optimization objective function is established to overcome false negative and positive artifacts induced by limited-angle measurements. We integrate neural network priors as a regularizer to explore deep features within residual data. Furthermore, the block-coordinate descent is employed to achieve a novel iterative framework. Second, a convolution assisted transformer is carefully elaborated to capture both local and long-range pixel interactions simultaneously. Regarding the visual transformer, the multi-head attention is further redesigned to reduce computational costs and protect reconstructed image features. Third, based on the relative error convergence property of the convolution assisted transformer, a mathematical convergence analysis is also provided for our IRON. Both numerically simulated and clinically collected real cardiac datasets are employed to validate the effectiveness and advantages of the proposed IRON. The results show that IRON outperforms other state-of-the-art methods.

Tomographic image reconstruction with deep learning is an emerging field of applied artificial intelligence. Reducing radiation dose with sparse views reconstruction is a significant task in cardiac imaging. Many efforts are contributing to sparse-view tomography imaging, but it is still a challenge for achieving good images from high sparse-view level, such as 60 views. In this study, we proposed a Deep Embedding-Attention-Refinement (DEAR) network to fundamentally address this challenge. DEAR consists of three modules including deep embedding, deep attention and deep refinement. The measurement is extended by deep embedding network to generate artifact-reduction images. Then, the deep attention network is employed to remove sparse-view artifacts and correct wrong details introduced by deep embedding network. Finally, the deep refinement module is used to refine finer image features and structures. The results on clinical datasets demonstrate the efficiency of our proposed DEAR in edge preservation and feature recovery.

Photon-counting X-ray computed tomography (CT) has been attracting great attention in tissue characterization, material discrimination, and so on. The emitting X-ray energy spectrum cutting into several energy bins that can result in only a part of X-ray photons can be collected within each narrow bin. This can compromise the image quality. In this case, how to obtain high-quality tomography is a big challenge. In this study, to overcome these issues, we mainly focus on developing an advanced imaging software based on the latest photon-counting CT system (MARS scanner). Specifically, we first design a weight adaptive total variation (TV) using compressed sensing theory. Then, combining the weight adaptive TV and nonlocal low-rank tensor factorization to formulate a new weight adaptive total-variation and image-spectral tensor factorization (WATITF) model for high-quality imaging. Finally, the optimization model is performed to obtain its solution. The studies including the numerical and preclinical mice are performed to validate and evaluate its outperformance.

Score-based generative model (SGM) has risen to prominence in sparse-view CT reconstruction due to its impressive generation capability. The consistency of data is crucial in guiding the reconstruction process in SGM-based reconstruction methods. However, the existing data consistency policy exhibits certain limitations. Firstly, it employs partial data from the reconstructed image of iteration process for image updates, which leads to secondary artifacts with compromising image quality. Moreover, the updates to the SGM and data consistency are considered as distinct stages, disregarding their interdependent relationship. Additionally, the reference image used to compute gradients in the reconstruction process is derived from intermediate result rather than ground truth. Motivated by the fact that a typical SGM yields distinct outcomes with different random noise inputs, we propose a Multi-channel Optimization Generative Model (MOGM) for stable ultra-sparse-view CT reconstruction by integrating a novel data consistency term into the stochastic differential equation model. Notably, the unique aspect of this data consistency component is its exclusive reliance on original data for effectively confining generation outcomes. Furthermore, we pioneer an inference strategy that traces back from the current iteration result to ground truth, enhancing reconstruction stability through foundational theoretical support. We also establish a multi-channel optimization reconstruction framework, where conventional iterative techniques are employed to seek the reconstruction solution. Quantitative and qualitative assessments on 23 views datasets from numerical simulation, clinical cardiac and sheep’s lung underscore the superiority of MOGM over alternative methods. Reconstructing from just 10 and 7 views, our method consistently demonstrates exceptional performance.

Spectral-computed tomography (CT) has been demonstrating its great advantages in lesion detection, tissue characterization, and material decomposition. However, the quality of images is often significantly corrupted with various noises, which brings a great challenge for its applications. Because the channel-wise images from different energy interval share similar structure and physical message, the spatial sparsity, global correlation across the spectrum (GCS), and nonlocal self-similarity (NSS) as three important characteristics are employed to spectral CT reconstruction. In this study, we propose a spectral-image decomposition with energy-fusion sensing (SIDES) reconstruction method, which encourages to obtain better quality spectral images and material decomposition results by establishing a unified tensor decomposition model. First, considering the noise distribution in channel-wise and the difference of linear attenuation coefficients within channel-cross, an adaptive weighted full-spectrum prior image as additional supervised information is incorporated to formulate a new weighted prior image-based tensor. Cooperating with original image tensor, they fully explore the spatial sparsity, GCS, and NSS properties. Then, we formulate nonlocal similar patch-based tensor groups to encode the NSS property from image-domain and residual-image-domain (which is expanded by prior-image and image-self). Next, low-rank regularized Tucker tensor decomposition is employed to fully explore the intrinsic knowledge with the help of prior-image supervision. Finally, the relaxed convex optimization model is optimized by dividing reconstruction model into several subproblem using split-Bregman method. Numerical simulations and real experiments are designed to validate and evaluate the SIDES method and the results demonstrate that the SIDES reconstruction outperforms the state-of-the-art.

The spectral computed tomography (CT) system based on a photon-counting detector (PCD) can quantitatively analyze the material composition of the inspected object by material decomposition. Nonetheless, the raw projection of spectral CT is frequently disturbed by noise and artifacts, resulting in poor quality material decomposition images. Recently, a generalized dictionary learning (DL) regularization was introduced into the image-domain material decomposition (GDLIMD) to obtain high-quality material images. DL has great advantages in noise suppression and artifacts, while its protection of the fine structure and edge information is insufficient. To address this limitation, we proposed a Sparsity Residual Prior and Dictionary Learning (SRPDL) algorithm for spectral CT image-domain material decomposition. The SRPDL method retains the noise-resistance performance of DL while introducing the pixel-value-based l ₀ norm constraint to guide the material decomposition process by using the structural redundancy information between the prior image and the material images, which further improves structure protection and reduces material misclassification. We conducted numerical simulations, physical phantom, and preclinical experiments to validate and evaluate the SRPDL method. The results demonstrate that the proposed SRPDL method obtained better material decomposition accuracy than the state-of-the-art methods in noise reduction and edge protection.