A Denoising Framework for Real-World Ultra-Low-Dose Lung CT Images Based on an Image Purification Strategy

Chest X-ray radiography (CXR) is a vital imaging technique for diagnosing diseases, but its 2D nature limits the capture of 3D anatomical structures. To overcome this, a new model named X-WIN has been developed, which predicts 2D projections from 3D chest computed tomography (CT) data. This model utilizes an affinity-guided contrastive alignment loss to enhance the learning of correlated information across projections, aiming to improve disease diagnosis and representation learning.

Computed tomography (CT) is a crucial imaging technique used in various fields, but it faces challenges due to limited scanning angles, resulting in distorted images. This study introduces Residual Null-Space Diffusion Stochastic Differential Equations (RN-SDEs) to tackle the Limited Angle Computed Tomography (LACT) reconstruction problem. Experiments on two datasets, ChromSTEM and C4KC-KiTS, demonstrate the effectiveness of RN-SDEs in improving image quality by leveraging learned mean-reverting stochastic differential equations.

The Poisson-Conditioned Generative Model (PoCGM) has been proposed to enhance sparse-view computed tomography (CT) reconstruction. This model aims to mitigate the challenges of aliasing artifacts and loss of structural details that arise from reducing projection views, which is crucial for minimizing radiation exposure and improving temporal resolution. By reformulating the Poisson Flow Generative Model (PFGM++) into a conditional framework, PoCGM integrates sparse-view data during training and sampling, improving the quality of reconstructed images.

The paper introduces Knowledge-Informed Dynamic Optimal Transport (KIDOT), a framework aimed at improving medical image reconstruction from measurement data. Traditional deep learning methods often rely on paired measurements and high-quality images, leading to performance issues when applied to real prospective data due to the retrospective-to-prospective gap. KIDOT addresses this by conceptualizing reconstruction as a dynamic transport path, learning from unpaired data and ensuring consistency with imaging physics.