3D Photon Counting CT Image Super-Resolution Using Conditional Diffusion Model
Authors:
Chuang Niu,
Christopher Wiedeman,
Mengzhou Li,
Jonathan S Maltz,
Ge Wang
Abstract:
This study aims to improve photon counting CT (PCCT) image resolution using denoising diffusion probabilistic models (DDPM). Although DDPMs have shown superior performance when applied to various computer vision tasks, their effectiveness has yet to be translated to high dimensional CT super-resolution. To train DDPMs in a conditional sampling manner, we first leverage CatSim to simulate realistic…
▽ More
This study aims to improve photon counting CT (PCCT) image resolution using denoising diffusion probabilistic models (DDPM). Although DDPMs have shown superior performance when applied to various computer vision tasks, their effectiveness has yet to be translated to high dimensional CT super-resolution. To train DDPMs in a conditional sampling manner, we first leverage CatSim to simulate realistic lower resolution PCCT images from high-resolution CT scans. Since maximizing DDPM performance is time-consuming for both inference and training, especially on high-dimensional PCCT data, we explore both 2D and 3D networks for conditional DDPM and apply methods to accelerate training. In particular, we decompose the 3D task into efficient 2D DDPMs and design a joint 2D inference in the reverse diffusion process that synergizes 2D results of all three dimensions to make the final 3D prediction. Experimental results show that our DDPM achieves improved results versus baseline reference models in recovering high-frequency structures, suggesting that a framework based on realistic simulation and DDPM shows promise for improving PCCT resolution.
△ Less
Submitted 21 August, 2024;
originally announced August 2024.
PPFM: Image denoising in photon-counting CT using single-step posterior sampling Poisson flow generative models
Authors:
Dennis Hein,
Staffan Holmin,
Timothy Szczykutowicz,
Jonathan S Maltz,
Mats Danielsson,
Ge Wang,
Mats Persson
Abstract:
Diffusion and Poisson flow models have shown impressive performance in a wide range of generative tasks, including low-dose CT image denoising. However, one limitation in general, and for clinical applications in particular, is slow sampling. Due to their iterative nature, the number of function evaluations (NFE) required is usually on the order of $10-10^3$, both for conditional and unconditional…
▽ More
Diffusion and Poisson flow models have shown impressive performance in a wide range of generative tasks, including low-dose CT image denoising. However, one limitation in general, and for clinical applications in particular, is slow sampling. Due to their iterative nature, the number of function evaluations (NFE) required is usually on the order of $10-10^3$, both for conditional and unconditional generation. In this paper, we present posterior sampling Poisson flow generative models (PPFM), a novel image denoising technique for low-dose and photon-counting CT that produces excellent image quality whilst keeping NFE=1. Updating the training and sampling processes of Poisson flow generative models (PFGM)++, we learn a conditional generator which defines a trajectory between the prior noise distribution and the posterior distribution of interest. We additionally hijack and regularize the sampling process to achieve NFE=1. Our results shed light on the benefits of the PFGM++ framework compared to diffusion models. In addition, PPFM is shown to perform favorably compared to current state-of-the-art diffusion-style models with NFE=1, consistency models, as well as popular deep learning and non-deep learning-based image denoising techniques, on clinical low-dose CT images and clinical images from a prototype photon-counting CT system.
△ Less
Submitted 19 December, 2023; v1 submitted 15 December, 2023;
originally announced December 2023.
