Open AccessDeformable registration of chest CT images using a 3D convolutional neural network based on unsupervised learning
Author(s) -
Zheng Yongnan,
Jiang Shan,
Yang Zhiyong
Publication year - 2021
Publication title -
journal of applied clinical medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.83
H-Index - 48
ISSN - 1526-9914
DOI - 10.1002/acm2.13392
Subject(s) - artificial intelligence , image registration , computer science , voxel , computer vision , displacement field , deep learning , convolutional neural network , regularization (linguistics) , artificial neural network , medical imaging , overfitting , euclidean vector , pattern recognition (psychology) , mathematics , image (mathematics) , physics , geometry , finite element method , thermodynamics
Abstract Purpose The deformable registration of 3D chest computed tomography (CT) images is one of the most important tasks in the field of medical image registration. However, the nonlinear deformation and large‐scale displacement of lung tissues caused by respiratory motion cause great challenges in the deformable registration of 3D lung CT images. Materials and methods We proposed an end‐to‐end fast registration method based on unsupervised learning, optimized the classic U‐Net, and added inception modules between skip connections. The inception module attempts to capture and merge information at different spatial scales to generate a high‐precision dense displacement vector field. To solve the problem of voxel folding in flexible registration, we put the Jacobian regularization term into the loss function to directly penalize the singularity of the displacement field during training to ensure a smooth displacement vector field. In the stage of data preprocessing, we segmented the lung fields to eliminate the interference of irrelevant information in the network during training. The existing publicly available datasets cannot implement model training. To alleviate the problem of overfitting caused by limited data resources being available, we proposed a data augmentation method based on the 3D‐TPS (3D thin plate spline) transform to expand the training data. Results Compared with the experimental results obtained by using the VoxelMorph deep learning method and registration packages, such as ANTs and Elastix, we achieved a competitive target registration error of 2.09 mm, an optimal Dice score of 0.987, and almost no folding voxels. Additionally, the proposed method was much faster than the traditional methods. Conclusions In this study, we have shown that the proposed method was efficient in 3D chest image registration. The promising results demonstrated that our method showed strong robustness in the deformable registration of 3D chest CT images.