AST-n: A Fast Sampling Approach for Low-Dose CT Reconstruction using Diffusion Models

TL;DR

AST-n introduces an accelerated diffusion-based framework for low-dose CT reconstruction, significantly reducing inference time while maintaining high image quality, thus enhancing clinical applicability.

Contribution

The paper presents AST-n, a novel accelerated inference method for diffusion models that enables rapid low-dose CT image reconstruction with minimal quality loss.

Findings

AST-n achieves high PSNR and SSIM with only 25 steps.

Inference time is reduced from ~16 seconds to under 1 second per slice.

Conditioned models outperform unconditional sampling in image quality.

Abstract

Low-dose CT (LDCT) protocols reduce radiation exposure but increase image noise, compromising diagnostic confidence. Diffusion-based generative models have shown promise for LDCT denoising by learning image priors and performing iterative refinement. In this work, we introduce AST-n, an accelerated inference framework that initiates reverse diffusion from intermediate noise levels, and integrate high-order ODE solvers within conditioned models to further reduce sampling steps. We evaluate two acceleration paradigms--AST-n sampling and standard scheduling with high-order solvers -- on the Low Dose CT Grand Challenge dataset, covering head, abdominal, and chest scans at 10-25 % of standard dose. Conditioned models using only 25 steps (AST-25) achieve peak signal-to-noise ratio (PSNR) above 38 dB and structural similarity index (SSIM) above 0.95, closely matching standard baselines while cutting inference time from ~16 seg to under 1 seg per slice. Unconditional sampling suffers substantial quality loss, underscoring the necessity of conditioning. We also assess DDIM inversion, which yields marginal PSNR gains at the cost of doubling inference time, limiting its clinical practicality. Our results demonstrate that AST-n with high-order samplers enables rapid LDCT reconstruction without significant loss of image fidelity, advancing the feasibility of diffusion-based methods in clinical workflows.