A novel strategy to develop deep learning for image super-resolution using original ultra-high-resolution computed tomography images of lung as training dataset

被引:15
作者
Kitahara, Hitoshi [1 ]
Nagatani, Yukihiro [1 ]
Otani, Hideji [1 ]
Nakayama, Ryohei [2 ]
Kida, Yukako [1 ]
Sonoda, Akinaga [1 ]
Watanabe, Yoshiyuki [1 ]
机构
[1] Shiga Univ Med Sci, Dept Radiol, Seta Tsukinowa Cho, Otsu, Shiga 5202192, Japan
[2] Ritsumeikan Univ, Dept Elect & Comp Engn, 1-1-1 Noji Higashi, Kusatsu, Shiga 5258577, Japan
关键词
Deep learning; Artificial intelligence; Convolutional neural networks; Cadaveric lung; Ultra-high-resolution computed tomography; CT;
D O I
10.1007/s11604-021-01184-8
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
Purpose To improve the image quality of inflated fixed cadaveric human lungs by utilizing ultra-high-resolution computed tomography (U-HRCT) as a training dataset for super-resolution processing using deep learning (SR-DL). Materials and methods Image data of nine cadaveric human lungs were acquired using U-HRCT. Three different matrix images of U-HRCT images were obtained with two acquisition modes: normal mode (512-matrix image) and super-high-resolution mode (1024- and 2048-matrix image). SR-DL used 512- and 1024-matrix images as training data for deep learning. The virtual 2048-matrix images were acquired by applying SR-DL to the 1024-matrix images. Three independent observers scored normal anatomical structures and abnormal computed tomography (CT) findings of both types of 2048-matrix images on a 3-point scale compared to 1024-matrix images. The image noise values were quantitatively calculated. Moreover, the edge rise distance (ERD) and edge rise slope (ERS) were also calculated using the CT attenuation profile to evaluate margin sharpness. Results The virtual 2048-matrix images significantly improved visualization of normal anatomical structures and abnormal CT findings, except for consolidation and nodules, compared with the conventional 2048-matrix images (p < 0.01). Quantitative noise values were significantly lower in the virtual 2048-matrix images than in the conventional 2048-matrix images (p < 0.001). ERD was significantly shorter in the virtual 2048-matrix images than in the conventional 2048-matrix images (p < 0.01). ERS was significantly higher in the virtual 2048-matrix images than in the conventional 2048-matrix images (p < 0.01). Conclusion SR-DL using original U-HRCT images as a training dataset might be a promising tool for image enhancement in terms of margin sharpness and image noise reduction. By applying trained SR-DL to U-HRCT SHR mode images, virtual ultra-high-resolution images were obtained which surpassed the image quality of unmodified SHR mode images.
引用
收藏
页码:38 / 47
页数:10
相关论文
共 50 条
  • [41] Edge-aware and edge-enhancing single digital image super-resolution using deep learning
    Kolli, Abhiram
    De Silva, Perumadura
    Mohsenzadegan, Kabeh
    Tavakkoli, Vahid
    Al Sayed, Mohamad
    Kyamakya, Kyandoghere
    Pichler, Ralf
    Bouwmeester, Olaf
    Zupan, Robert
    [J]. DEVELOPMENTS OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES IN COMPUTATION AND ROBOTICS, 2020, 12 : 1069 - 1077
  • [42] Rock core micrograph image super-resolution reconstruction using a novel deep residual network
    Huo, Fengcai
    Liu, Kaiming
    Dong, Hongli
    Ren, Weijian
    Zhang, Heng
    [J]. SIGNAL IMAGE AND VIDEO PROCESSING, 2025, 19 (05)
  • [43] Ultra-high-resolution computed tomography can demonstrate alveolar collapse in novel coronavirus (COVID-19) pneumonia
    Tae Iwasawa
    Midori Sato
    Takafumi Yamaya
    Yozo Sato
    Yoshinori Uchida
    Hideya Kitamura
    Eri Hagiwara
    Shigeru Komatsu
    Daisuke Utsunomiya
    Takashi Ogura
    [J]. Japanese Journal of Radiology, 2020, 38 : 394 - 398
  • [44] A Novel Deep-Learning-Based Enhanced Texture Transformer Network for Reference Image Super-Resolution
    Liu, Changhong
    Li, Hongyin
    Liang, Zhongwei
    Zhang, Yongjun
    Yan, Yier
    Zhong, Ray Y.
    Peng, Shaohu
    [J]. ELECTRONICS, 2022, 11 (19)
  • [45] Assessment of diagnostic image quality of computed tomography (CT) images of the lung using deep learning
    Lee, John H.
    Grant, Byron R.
    Chung, Jonathan H.
    Reiser, Ingrid
    Giger, Maryellen
    [J]. MEDICAL IMAGING 2018: PHYSICS OF MEDICAL IMAGING, 2018, 10573
  • [46] Investigating single image super-resolution algorithm with deep learning using convolutional neural network for chest digital tomosynthesis
    Kim, Tae-Hyung
    Oh, Heekyung
    Kim, Kyuseok
    Lee, Youngjin
    [J]. OPTIK, 2020, 203
  • [47] Convolutional Neural Networks Using Skip Connections with Layer Groups for Super-Resolution Image Reconstruction Based on Deep Learning
    Ahn, Hyeongyeom
    Yim, Changhoon
    [J]. APPLIED SCIENCES-BASEL, 2020, 10 (06):
  • [48] High-Throughput Deep Learning Microscopy Using Multi-Angle Super-Resolution
    Zhang, Jizhou
    Xu, Tingfa
    Li, Xiangmin
    Zhang, Yizhou
    Chen, Yiwen
    Wang, Xin
    Wang, Shushan
    Wang, Chen
    [J]. IEEE PHOTONICS JOURNAL, 2020, 12 (02):
  • [49] Deep Unpaired Blind Image Super-Resolution Using Self-supervised Learning and Exemplar Distillation
    Dong, Jiangxin
    Bai, Haoran
    Tang, Jinhui
    Pan, Jinshan
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2023,
  • [50] Ultra-high-resolution photon-counting detector computed tomography of the lungs: Phantom and clinical assessment of radiation dose and image quality
    Chamberlin, Jordan H.
    Smith, Carter D.
    Maisuria, Dhruw
    Parrish, Joe
    van Swol, Elizabeth
    Mah, Eugene
    Emrich, Tilman
    Schoepf, U. Joseph
    Varga-Szemes, Akos
    O'Doherty, Jim
    Munden, Reginald F.
    Tipnis, Sameer, V
    Baruah, Dhiraj
    Kabakus, Ismail M.
    [J]. CLINICAL IMAGING, 2023, 104