Low-dose CT via convolutional neural network

被引：740

作者：

Chen, Hu ^{[1
,2
]}

Zhang, Yi ^{[1
]}

Zhang, Weihua ^{[1
]}

Liao, Peixi ^{[3
]}

Li, Ke ^{[1
,2
]}

Zhou, Jiliu ^{[1
]}

Wang, Ge ^{[4
]}

机构：

[1] Sichuan Univ, Coll Comp Sci, Chengdu 610065, Peoples R China

[2] Sichuan Univ, Natl Key Lab Fundamental Sci Synthet Vis, Chengdu 610065, Peoples R China

[3] Sixth Peoples Hosp Chengdu, Dept Sci Res & Educ, Chengdu 610065, Peoples R China

[4] Rensselaer Polytech Inst, Dept Biomed Engn, Troy, NY 12180 USA

来源：

BIOMEDICAL OPTICS EXPRESS | 2017年 / 8卷 / 02期

基金：

美国国家卫生研究院; 中国国家自然科学基金;

关键词：

VIEW IMAGE-RECONSTRUCTION; WEIGHTED LEAST-SQUARES; COMPUTED-TOMOGRAPHY; ALGORITHM; SPARSE; CANCER; SEGMENTATION; RESTORATION; REDUCTION; NUCLEI;

D O I：

10.1364/BOE.8.000679

中图分类号：

Q5 [生物化学];

学科分类号：

070307 [化学生物学];

摘要：

In order to reduce the potential radiation risk, low-dose CT has attracted an increasing attention. However, simply lowering the radiation dose will significantly degrade the image quality. In this paper, we propose a new noise reduction method for low-dose CT via deep learning without accessing original projection data. A deep convolutional neural network is here used to map low-dose CT images towards its corresponding normal-dose counterparts in a patch-by-patch fashion. Qualitative results demonstrate a great potential of the proposed method on artifact reduction and structure preservation. In terms of the quantitative metrics, the proposed method has showed a substantial improvement on PSNR, RMSE and SSIM than the competing state-of-art methods. Furthermore, the speed of our method is one order of magnitude faster than the iterative reconstruction and patch-based image denoising methods. (C) 2017 Optical Society of America

引用

页码：679 / 694

页数：16

共 53 条

[1]

K-SVD: An algorithm for designing overcomplete dictionaries for sparse representation [J].