Nonlinear CT Image Reconstruction

Nonlinear CT image reconstruction aims to overcome limitations of traditional linear methods by directly addressing the inherently nonlinear physics of X-ray attenuation in computed tomography. Current research focuses on developing and analyzing novel algorithms, including diffusion models, gradient descent methods, and neural networks (like Polyner and variations of DeepONet), to solve this challenging inverse problem. These advancements improve image quality, particularly in reducing artifacts near high-density materials like metal, and enable more accurate quantification of features of interest, even with limited data. The resulting improvements have significant implications for medical diagnosis and other applications requiring high-fidelity CT imaging.