China
Africa
Explore chapters and articles related to this topic
Cardiac Fiber Imaging with 3D Ultrasound and MR Diffusion Tensor Imaging
Published in Ayman El-Baz, Jasjit S. Suri, Cardiovascular Imaging and Image Analysis, 2018
To this end, a new pipeline has been proposed similarly based on the relationship between fiber orientations and geometry of the heart. But different from the rule-based methods, it uses cardiac fiber atlas of ex vivo DTI as the template to provide patient-specific cardiac fiber orientations [25] and has been applied in cardiac diagnosis and therapies [32]. Generally, this approach estimates cardiac fiber orientations from the template DTI based on geometric similarity between both target and template hearts. This similarity is mostly measured by registering both volumetric geometries. Following this approach, Helm et al. proposed an algorithm of large deformation diffeomorphic metric mapping (LDDMM) to map cardiac fiber orientations from ex vivo DTI data to the target heart. After mapping, they further validated the results by comparing with their histological findings [33]. Later, a new registration approach of elastic registration was proposed to map the DTI template onto a patient-specific heart [34]. Similarly, using Demons registration, Zhang et al. provided an atlas-based geometry pipeline that could deform DTI data to patient-specific cardiac geometries for constructing three-dimensional cubic Hermite finite element meshes of the whole human heart [15]. Moreover, to validate the accuracy of this geometry-based approach for electrophysiological simulations, Vadakkumpadan et al. used LDDMM to register the MRI geometry of an ex vivo heart atlas to the CT geometry of an in vivo target heart and then deformed the diffusion tensors of the atlas as the estimated fiber orientations of the patient [25]. Their simulation results demonstrated that the estimated fiber orientations through this approach only slightly affected the electrophysiological properties of the target heart.
Deformable Registration Choices for Multi-Atlas Segmentation
Published in Jinzhong Yang, Gregory C. Sharp, Mark J. Gooding, Auto-Segmentation for Radiation Oncology, 2021
Keyur Shah, James Shackleford, Nagarajan Kandasamy, Gregory C. Sharp
Several researchers have explored image registration approaches for atlas-based segmentation. Alven et al. presented a feature-based registration method which combines the information of the entire atlas set and efficiently finds robust correspondences and transformations between the target and all the images in the atlas set [1]. Bai et. al. compared, in their work, the performance of four different image registration algorithms: affine, B-spline, free-form deformation, and large deformation diffeomorphic metric mapping (LDDMM). Their experiments found that the LDDMM registration algorithm worked best for the mouse brain image segmentation task [2]. Datteri et al. used the adaptive bases algorithm (ABA) with an NMI objective function for registration. ABA models the deformation field as a linear combination of radial basis functions with finite support [3]. Doshi et al. used the Advanced Normalization Tools (ANTs) registration toolkit for atlas registration [4]. Heckemann et al. compare the image registration toolkit (IRTK) [5], based on B-splines and maximizing the NMI, with an algorithm they termed multi-atlas propagation with enhanced registration (MAPER), which optimizes both image intensity and tissue classification. They found that the MAPER approach provides superior results for a brain segmentation task [6]. Lötjönen et al. introduced a similarity metric based on intensity normalized images and compared it with NMI. They found a threefold reduction in the computation time with similar registration accuracy [7]. Sjöberg et al. compared a B-spline based method with a demons algorithm for registration strategy. No significant differences were reported [8]. Yeo et al. employed a generative model for the construction of a probabilistic atlas for joint registration and segmentation of images [9].
Shape-constrained Gaussian process regression for surface reconstruction and multimodal, non-rigid image registration
Published in Journal of Applied Statistics, 2022
Thomas Deregnaucourt, Chafik Samir, Sebastian Kurtek, Anne-Francoise Yao
For large deformations, a diffeomorphic matching approach was developed by [14]; it was followed by other deformable template-based approaches such as the Large Deformation Diffeomorphic Metric Mapping (LDDMM) methods [11,15] to solve for large deformations when landmark correspondences are known [5]. Other approaches can be found in computer vision or medical imaging, where the deformation fields are defined on smooth regions with distinct ‘anatomical’ landmark points. In this respect, geometric regularization and stochastic formulations offer some nice advantages [35]. Specifically, deformable image registration can be formulated as a stochastic optimization problem, in which the likelihood term is coupled with regularization priors to ensure smooth solutions. In this work, geometric landmarks are given as 2D curves without any correspondences between points (correspondences are only given at the curve level). In this setting, previous methods are not directly applicable, and one has to develop a unified framework that can efficiently compute point correspondences across curves and the full deformation fields jointly. This motivates us to focus on curve-based image registration, and to develop a novel probabilistic model that encodes shape variability of the landmark curves.
Automated CT bone segmentation using statistical shape modelling and local template matching
Published in Computer Methods in Biomechanics and Biomedical Engineering, 2019
Elham Taghizadeh, Alexandre Terrier, Fabio Becce, Alain Farron, Philippe Büchler
For these samples, the adaptation of coherent point drift proposed by Mutsvangwa et al. (2015) was used to register all samples in the dataset to the reference bone. For mandibular bones, higher registration accuracy was achieved using the open-source software Deformetrica (http://www.deformetrica.org), which relies on the large deformation diffeomorphic metric mapping framework introduced in (Durrleman et al. 2014).