Unsupervised Denoising of Real Clinical Low Dose Liver CT with Perceptual Attention Networks
2026-05-01 • Artificial Intelligence
Artificial IntelligenceComputer Vision and Pattern Recognition
AI summaryⓘ
The authors focus on reducing noise in low-dose CT scans, which are safer but usually grainier and harder for doctors to read. They created a new deep learning method that doesn't need labeled data, combining U-Net, attention, and residual networks to clean up the images. To make sure their method works well, they tested it on real clinical CT scans and had doctors check the results. Their approach helps improve image quality without needing lots of perfectly matched data, which is often hard to get in hospitals.
low-dose computed tomographyimage denoisingdeep learningCycle-GANU-Netattention mechanismresidual networkperceptual lossunsupervised learningmedical image processing
Authors
Jingxi Pu, Tonghua Liu, Zhilin Guan, Siqiao Li, Yang Ming, Zheng Cong, Wei Zhang, Fangwei Li
Abstract
With the development of deep learning, medical image processing has been widely used to assist clinical research. This paper focuses on the denoising problem of low-dose computed tomography using deep learning. Although low-dose computed tomography reduces radiation exposure to patients, it also introduces more noise, which may interfere with visual interpretation by physicians and affect diagnostic results. To address this problem, inspired by Cycle-GAN for unsupervised learning, this paper proposes an end-to-end unsupervised low-dose computed tomography denoising framework. The proposed framework combines a U-Net structure for multi-scale feature extraction, an attention mechanism for feature fusion, and a residual network for feature transformation. It also introduces perceptual loss to improve the network for the characteristics of medical images. In addition, we construct a real low-dose computed tomography dataset and design a large number of comparative experiments to validate the proposed method, using both image-based evaluation metrics and medical evaluation criteria. Compared with classical methods, the main advantage of this paper is that it addresses the limitation that real clinical data cannot be directly used for supervised learning, while still achieving excellent performance. The experimental results are also professionally evaluated by imaging physicians and meet clinical needs.