Research progress in low-dose computed tomography image denoising algorithm based on deep learning
10.3760/cma.j.cn121382-20240829-00614
- VernacularTitle:基于深度学习的低剂量计算机断层扫描图像去噪算法的研究进展
- Author:
Shengjun ZHANG
1
;
Chunxiang WANG
Author Information
1. 天津市儿童医院（天津大学儿童医院）医学影像科，天津 300074
- Keywords:
Computed tomography;
Low-dose;
Deep learning;
Image denoising
- From:
International Journal of Biomedical Engineering
2024;47(6):616-621
- CountryChina
- Language:Chinese
-
Abstract:
Higher doses of radiation from computed tomography (CT) scans can cause health problems. The radiation dose should be as low as possible while still meeting diagnostic needs. However, low-dose computed tomography (LDCT) increases image noise, reduces image quality and compromises diagnostic accuracy. Researchers have focused on the denoising technology for LDCT image. In this review, the research progress of projection domain pre-processing denoising, image reconstruction denoising, image domain post-processing denoising and dual-domain processing denoising algorithm based on deep learning for LDCT image were reviewed. The current problems of LDCT denoising were also summarized to promote the application of deep learning in LDCT image denoising.
