Deep neural networks-based denoising models for CT imaging and their efficacy

被引:9
|
作者
Kc, Prabhat [1 ]
Zeng, Rongping [1 ]
Farhangi, M. Mehdi [1 ]
Myers, Kyle J. [1 ]
机构
[1] US FDA, Silver Spring, MD 20993 USA
来源
MEDICAL IMAGING 2021: PHYSICS OF MEDICAL IMAGING | 2021年 / 11595卷
关键词
CT image denoising; deep learning; neural networks; loss functions; image quality; LOW-DOSE CT; NUMBERS;
D O I
10.1117/12.2581418
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Most of the Deep Neural Networks (DNNs) based CT image denoising literature shows that DNNs outperform traditional iterative methods in terms of metrics such as the RMSE, the PSNR and the SSIM. In many instances, using the same metrics, the DNN results from low-dose inputs are also shown to be comparable to their high-dose counterparts. However, these metrics do not reveal if the DNN results preserve the visibility of subtle lesions or if they alter the CT image properties such as the noise texture. Accordingly, in this work, we seek to examine the image quality of the DNN results from a holistic viewpoint for low-dose CT image denoising. First, we build a library of advanced DNN denoising architectures. This library is comprised of denoising architectures such as the DnCNN, U-Net, Red-Net, GAN, etc. Next, each network is modeled, as well as trained, such that it yields its best performance in terms of the PSNR and SSIM. As such, data inputs (e.g. training patch-size, reconstruction kernel) and numeric-optimizer inputs (e.g. minibatch size, learning rate, loss function) are accordingly tuned. Finally, outputs from thus trained networks are further subjected to a series of CT bench testing metrics such as the contrast-dependent MTF, the NPS and the HU accuracy. These metrics are employed to perform a more nuanced study of the resolution of the DNN outputs' low-contrast features, their noise textures, and their CT number accuracy to better understand the impact each DNN algorithm has on these underlying attributes of image quality.
引用
收藏
页数:13
相关论文
共 50 条
  • [41] Jacobian Analysis in Deep Learning CT Denoising Networks
    Zhang, Xiaoxuan
    Li, Patrick
    Li, Junyuan
    Gang, Grace Jianan
    MEDICAL IMAGING 2024: PHYSICS OF MEDICAL IMAGING, PT 1, 2024, 12925
  • [42] Spatial Information and Deep Neural Networks-Based Smart Design for Light Environment of Museum Exhibition
    Yang, Guojie
    He, Honghao
    JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS, 2024, 33 (05)
  • [43] Open-Set Recognition for Deep Neural Networks-based Underwater Acoustic Target Classification
    Nie, Leixin
    Li, Chao
    Wang, Haibin
    Marzani, Franck
    OCEANS 2021: SAN DIEGO - PORTO, 2021,
  • [44] Application of Neural Networks to Denoising of CT Images of Lungs
    Turajlic, Emir
    2016 XI INTERNATIONAL SYMPOSIUM ON TELECOMMUNICATIONS (BIHTEL), 2016,
  • [45] An Intelligent Deep Convolutional Neural Networks-Based Islanding Detection for Multi-DG Systems
    Hussain, Arif
    Kim, Chul-Hwan
    Jabbar, Muhammad Shahid
    IEEE ACCESS, 2022, 10 : 131920 - 131931
  • [46] Stability Analysis of Neural Networks-Based System Identification
    Korkobi, Talel
    Djemel, Mohamed
    Chtourou, Andmohamed
    MODELLING AND SIMULATION IN ENGINEERING, 2008, 2008
  • [47] DEEP SPARSE RECTIFIER NEURAL NETWORKS FOR SPEECH DENOISING
    Xu, Lie
    Choy, Chiu-Sing
    Li, Yi-Wen
    2016 IEEE INTERNATIONAL WORKSHOP ON ACOUSTIC SIGNAL ENHANCEMENT (IWAENC), 2016,
  • [48] Neural networks-based parametric testing of analog IC
    Stopjaková, V
    Micusík, D
    Benusková, L
    Margala, M
    17TH IEEE INTERNATIONAL SYMPOSIUM ON DEFECT AND FAULT TOLERANCE IN VLSI SYSTEMS, PROCEEDINGS, 2002, : 408 - 416
  • [49] Denoising wavefront sensor images with deep neural networks
    Pou, B.
    Quinones, E.
    Gratadour, D.
    Martin, M.
    ADAPTIVE OPTICS SYSTEMS VII, 2020, 11448
  • [50] Deep Parameterized Neural Networks for Hyperspectral Image Denoising
    Xiong, Fengchao
    Zhou, Jun
    Zhou, Jiantao
    Lu, Jianfeng
    Qian, Yuntao
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2023, 61