深度学习重建算法的原理及其在腹部CT临床应用进展  

Principles of Deep Learning Reconstruction Algorithm and Its Clinical Application Progress in Abdominal CT

作　　者：李云成 邓炜 李小虎[1] LI Yuncheng;DENG Wei;LI Xiaohu(Department of Radiology,the First Affiliated Hospital of Anhui Medical University,Heifei 230022,China)

机构地区：[1]安徽医科大学第一附属医院放射科,安徽合肥230022

出　　处：《中国医学影像学杂志》2025年第1期102-106,共5页Chinese Journal of Medical Imaging

摘　　要：深度学习重建(DLR)算法是近年新兴的CT图像重建技术,在临床实践中可作为滤波反投影和迭代重建的替代方案。相较于传统的图像重建算法,DLR算法能够在降低图像噪声和辐射剂量的同时,保留图像纹理,缩短重建时间,提高诊断效能,在图像重建领域具有广阔的临床应用前景。本文就DLR算法的基本原理及其在腹部CT临床应用新进展进行综述。The deep learning reconstruction(DLR)algorithm is a new CT image reconstruction technology in recent years,which can be used as an alternative to filtered back projection and iterative reconstruction in clinical practice.Compared with traditional image reconstruction algorithms,the DLR algorithm can reduce image noise and radiation dose while preserving image texture,shortening reconstruction time,and improving diagnostic efficiency.Therefore,it has broad clinical application prospects in the field of image reconstruction.This article summarizes the basic principles of the DLR algorithm and its new progress in the clinical application of abdominal CT.

关 键 词：体层摄影术 X线计算机 深度学习 图像重建算法 腹部 综述 

分 类 号：R445.3[医药卫生—影像医学与核医学] TP181[医药卫生—诊断学]