MAN: Latent Diffusion Enhanced Multistage Anti-Noise Network for Efficient and High-Quality Low-Dose CT Image Denoising

2025-09-27

AI Summary

To address the prohibitively long inference time (often exceeding 1,000 seconds) of diffusion models in low-dose CT (LDCT) image denoising—which hinders clinical deployment—this paper proposes a fast, deterministic denoising framework operating in a compressed latent space. The method comprises two key components: (i) a perception-optimized autoencoder that learns a high-fidelity, low-dimensional latent representation; and (ii) a lightweight, multi-stage conditional U-Net with integrated attention mechanisms, deployed in lieu of iterative diffusion sampling. By eliminating redundant noise prediction and sampling steps, the approach achieves denoising quality on par with heavyweight diffusion models such as DDPM (comparable PSNR and SSIM), while accelerating inference by over 60×. This substantial speedup significantly enhances clinical feasibility without compromising diagnostic image fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

While diffusion models have set a new benchmark for quality in Low-Dose Computed Tomography (LDCT) denoising, their clinical adoption is critically hindered by extreme computational costs, with inference times often exceeding thousands of seconds per scan. To overcome this barrier, we introduce MAN, a Latent Diffusion Enhanced Multistage Anti-Noise Network for Efficient and High-Quality Low-Dose CT Image Denoising task. Our method operates in a compressed latent space via a perceptually-optimized autoencoder, enabling an attention-based conditional U-Net to perform the fast, deterministic conditional denoising diffusion process with drastically reduced overhead. On the LDCT and Projection dataset, our model achieves superior perceptual quality, surpassing CNN/GAN-based methods while rivaling the reconstruction fidelity of computationally heavy diffusion models like DDPM and Dn-Dp. Most critically, in the inference stage, our model is over 60x faster than representative pixel space diffusion denoisers, while remaining competitive on PSNR/SSIM scores. By bridging the gap between high fidelity and clinical viability, our work demonstrates a practical path forward for advanced generative models in medical imaging.

Problem

• Reducing extreme computational costs in LDCT denoising
• Achieving high-quality image reconstruction with clinical viability
• Accelerating inference speed while maintaining competitive fidelity metrics

Innovation

• Latent diffusion in compressed space for efficiency
• Multistage anti-noise network with attention U-Net
• Fast deterministic denoising with perceptual optimization