Noise suppression in photon-counting CT using unsupervised Poisson flow generative models

TL;DR

This paper introduces a novel unsupervised Poisson flow generative model for photon-counting CT denoising, achieving high-quality results with a single-step sampling process that outperforms existing methods.

Contribution

The authors extend Poisson flow generative models to inverse problems in CT, enabling single-step denoising with improved robustness and performance on clinical data.

Findings

Achieves competitive denoising performance with NFE=1.

Outperforms supervised and other unsupervised methods on clinical CT data.

Provides a robust, efficient denoising technique for photon-counting CT.

Abstract

Deep learning has proven to be important for CT image denoising. However, such models are usually trained under supervision, requiring paired data that may be difficult to obtain in practice. Diffusion models offer unsupervised means of solving a wide range of inverse problems via posterior sampling. In particular, using the estimated unconditional score function of the prior distribution, obtained via unsupervised learning, one can sample from the desired posterior via hijacking and regularization. However, due to the iterative solvers used, the number of function evaluations (NFE) required may be orders of magnitudes larger than for single-step samplers. In this paper, we present a novel image denoising technique for photon-counting CT by extending the unsupervised approach to inverse problem solving to the case of Poisson flow generative models (PFGM)++. By hijacking and regularizing the sampling process we obtain a single-step sampler, that is NFE=1. Our proposed method incorporates posterior sampling using diffusion models as a special case. We demonstrate that the added robustness afforded by the PFGM++ framework yields significant performance gains. Our results indicate competitive performance compared to popular supervised, including state-of-the-art diffusion-style models with NFE=1 (consistency models), unsupervised, and non-deep learning-based image denoising techniques, on clinical low-dose CT data and clinical images from a prototype photon-counting CT system developed by GE HealthCare.