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 highprecision 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 Vox-elMorph 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.

Purpose To develop a method for automatically detecting needles from CT images, which can be used in image‐guided lung interstitial brachytherapy to assist needle placement assessment and dose distribution optimization. Material and Methods Based on the preview model parameters evaluation, local optimization combining local random sample consensus, and principal component analysis, the needle shaft was detected quickly, accurately, and robustly through the modified random sample consensus algorithm. By tracing intensities along the axis, the needle tip was determined. Furthermore, multineedles in a single slice were segmented at once using successive inliers deletion. Results The simulation data show that the segmentation efficiency is much higher than the original random sample consensus and yet maintains a stable submillimeter accuracy. Experiments with physical phantom demonstrate that the segmentation accuracy of described algorithm depends on the needle insertion depth into the CT image. Application to permanent lung brachytherapy image is also validated, where manual segmentation is the counterparts of the estimated needle shape. Conclusions From the results, the mean errors in determining needle orientation and endpoint are regulated within 2° and 1 mm, respectively. The average segmentation time is 0.238 s per needle.