Abstract
Deep-learning-based registration methods emerged as a fast alternative to conventional registration methods. However, these methods often still cannot achieve the same performance as conventional registration methods because they are either limited to small deformation or they fail to handle a superposition of large and small deformations without producing implausible deformation fields with foldings inside. In this paper, we identify important strategies of conventional registration methods for lung registration and successfully developed the deep-learning counterpart. We employ a Gaussian-pyramid-based multilevel framework that can solve the image registration optimization in a coarse-to-fine fashion. Furthermore, we prevent foldings of the deformation field and restrict the determinant of the Jacobian to physiologically meaningful values by combining a volume change penalty with a curvature regularizer in the loss function. Keypoint correspondences are integrated to focus on the alignment of smaller structures. We perform an extensive evaluation to assess the accuracy, the robustness, the plausibility of the estimated deformation fields, and the transferability of our registration approach. We show that it achieves state-of-the-art results on the COPDGene dataset compared to conventional registration method with much shorter execution time. In our experiments on the DIRLab exhale to inhale lung registration, we demonstrate substantial improvements (TRE below 1.2 mm) over other deep learning methods. Our algorithm is publicly available at https://grand-challenge.org/algorithms/deep-learning-based-ct-lung-registration/.
Highlights
Image registration is the process of aligning two or more images to achieve point-wise spatial correspondence
Our proposed method achieves on average significant better Dice Scores than the conventional registration method (0.95 vs. 0.92) with a smaller standard deviation (0.026 vs. 0.046)
Without the volume change control penalty, the deformation fields produced by our method show on average 0.30% of voxel positions with foldings, which is comparable to the values of other deep learning registration methods
Summary
Image registration is the process of aligning two or more images to achieve point-wise spatial correspondence. Image registration is phrased as an optimization problem with respect to a spatial mapping that minimizes a suitable cost function and common approaches estimate solutions by applying iterative optimization schemes. Solving such an optimization problem is computationally demanding and slow. While deep learning has become the methodology of choice in many areas, relatively few deep-learning-based image registration algorithms have been proposed One reason for this is the lack of ground truth and the large variability of plausible deformations that can align corresponding anatomies. As a trained neural networks can process images in real time, this has immense potential for time-sensitive applications such as image guidance in radiotherapy, tracking, or shape analysis through multi-atlas registration
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
