Infant Brain Age Classification: 2D CNN Outperforms 3D CNN in Small Dataset

被引:0
|
作者
Shabanian, Mahdieh [1 ]
Wenzel, Markus [2 ]
DeVincenzo, John P. [3 ]
机构
[1] Univ Tennessee, Ctr Hlth Sci, Dept Biomed Engn, Memphis, TN 38163 USA
[2] Fraunhofer Inst Digital Med MEVIS, Bremen, Germany
[3] Enanta Pharmaceut, Watertown, MA USA
来源
关键词
Deep learning; CNN; neurodevelopmental; age estimation; infant diseases; MACHINE; MRI;
D O I
10.1117/12.2612887
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
Determining if the brain is developing normally is a key component of pediatric neuroradiology and neurology. Brain magnetic resonance imaging (MRI) of infants demonstrates a specific pattern of development beyond simply myelination. While radiologists have used myelination patterns, brain morphology and size characteristics to determine age-adequate brain maturity, this requires years of experience in pediatric neuroradiology. With no standardized criteria, visual estimation of the structural maturity of the brain from MRI before three years of age remains dominated by inter-observer and intra-observer variability. A more objective estimation of brain developmental age could help physicians identify many neurodevelopmental conditions and diseases earlier and more reliably. Such data, however, is naturally hard to obtain, and the observer ground truth not much of a gold standard due to subjectivity of assessment. In this light, we explore the general feasibility to tackle this task, and the utility of different approaches, including two- and three-dimensional convolutional neural networks (CNN) that were trained on a fusion of T1-weighted, T2-weighted, and proton density (PD) weighted sequences from 84 individual subjects divided into four age groups from birth to 3 years of age. In the best performing approach, we achieved an accuracy of 0.90 [95% CI:0.86-0.94] using a 2D CNN on a central axial thick slab. We discuss the comparison to 3D networks and show how the performance compares to the use of only one sequence (T1w). In conclusion, despite the theoretical superiority of 3D CNN approaches, in limited-data situations, such approaches are inferior to simpler architectures. The code can be found in https://github.com/shabanian2018/Age_MRI-Classification
引用
收藏
页数:8
相关论文
共 50 条
  • [1] Hyperspectral image classification using 3D 2D CNN
    Diakite, Alou
    Gui, Jiangsheng
    Fu, Xiaping
    IET IMAGE PROCESSING, 2021, 15 (05) : 1083 - 1092
  • [2] Spatiotemporal Action Detection Using 2D CNN and 3D CNN
    Liu, Hengshuai
    Li, Jianjun
    Tang, Yuhong
    Zhang, Ningfei
    Zhang, Ming
    Wang, Yaping
    Li, Guang
    Computers and Electrical Engineering, 2024, 120
  • [3] Brain White Matter Lesion Segmentation with 2D/3D CNN
    Lopez-Zorrilla, A.
    de Velasco-Vazquez, M.
    Serradilla-Casado, O.
    Roa-Barco, L.
    Grana, M.
    Chyzhyk, D.
    Price, C. C.
    NATURAL AND ARTIFICIAL COMPUTATION FOR BIOMEDICINE AND NEUROSCIENCE, PT I, 2017, 10337 : 394 - 403
  • [4] A study on CNN image classification of EEG signals represented in 2D and 3D
    Bird, Jordan J.
    Faria, Diego R.
    Manso, Luis J.
    Ayrosa, Pedro P. S.
    Ekart, Aniko
    JOURNAL OF NEURAL ENGINEERING, 2021, 18 (02)
  • [5] MuellerNet: a hybrid 3D 2D CNN for cell classification with Mueller matrix images
    Ma, Dongdong
    Lu, Zongqing
    Xia, Longyu
    Liao, Qingmin
    Yang, Wenming
    Ma, Hui
    Liao, Ran
    Ma, Lan
    Liu, Zhidi
    APPLIED OPTICS, 2021, 60 (22) : 6682 - 6694
  • [6] 2D CNN versus 3D CNN for false-positive reduction in lung cancer screening
    Yu, Juezhao
    Yang, Bohan
    Wang, Jing
    Leader, Joseph
    Wilson, David
    Pu, Jiantao
    JOURNAL OF MEDICAL IMAGING, 2020, 7 (05)
  • [7] Fusion of 2D CNN and 3D DenseNet for Dynamic Gesture Recognition
    Zhang, Erhu
    Xue, Botao
    Cao, Fangzhou
    Duan, Jinghong
    Lin, Guangfeng
    Lei, Yifei
    ELECTRONICS, 2019, 8 (12)
  • [8] Hybrid Deep Feature Fusion of 2D CNN and 3D CNN for Vestibule Segmentation from CT Images
    Zhang, Ruicong
    Zhuo, Li
    Chen, Meijuan
    Yin, Hongxia
    Li, Xiaoguang
    Wang, Zhenchang
    COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE, 2022, 2022
  • [9] An Efficient and Robust 3D Medical Image Classification Approach Based on 3D CNN, Time-Distributed 2D CNN-BLSTM Models, and mRMR Feature Selection
    Akbacak, Enver
    Muzoğlu, Nedim
    Computational Intelligence, 2024, 40 (05)
  • [10] Combine EfficientNet and CNN for 3D model classification
    Gao, Xue-Yao
    Yang, Bo-Yu
    Zhang, Chun-Xiang
    MATHEMATICAL BIOSCIENCES AND ENGINEERING, 2023, 20 (05) : 9062 - 9079