Dense Displacement Vector Research Articles

An Unsupervised 3D Image Registration Network for Brain MRI Deformable Registration.

In recent years, deep learning has made successful applications and remarkable achievements in the field of medical image registration, and the method of medical image registration based on deep learning has become the current research hotspot. However, the performance of convolutional neural networks may not be fully exploited due to neglect of spatial relationships between distant locations in the image and incomplete updates of network parameters. To avoid this phenomenon, MHNet, a multiscale hierarchical deformable registration network for 3D brain MR images, was proposed in this paper. This network was an unsupervised end-to-end convolutional neural network. After training, the dense displacement vector field can be predicted almost in real-time for the unseen input image pairs, which saves a lot of time compared with the traditional algorithms of independent iterative optimization for each pair of images. On the basis of the encoder-decoder structure, this network introduced the improved Inception module for multiscale feature extraction and expanding the receptive field and the hierarchical forecast structure to promote the update of the parameters of the middle layers, which achieved the best performance on the augmented public dataset compared with the existing four excellent registration methods.

Deformable registration of chest CT images using a 3D convolutional neural network based on unsupervised learning.

PurposeThe 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 methodsWe 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.ResultsCompared 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.ConclusionsIn 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.

SU‐FF‐J‐57: Evaluating Lung Motion Variations in Repeated 4D CT Studies Using Inverse Consistent Image Registration

Purpose: To evaluate lung motion variations in repeated, amplitude‐based 4D CT studies using inverse consistent image registration. Method and Materials: Three patients with lung cancers were scanned twice with our 4D CT protocol. A 4D CT dataset was reconstructed using the amplitude and direction of an external respiratory signal. A thresholding‐based method was used to segment the lung. An inverse consistent image registration was used to estimate the intrafraction motion within each 4D study and the interfraction motion between two studies. This diffeomorphic registration jointly estimates the forward and reverse transformations providing more accurate correspondence between two images. We evaluated the lung motion variations in repeated 4D studies. Results: On average, the mean respiratory period changed by 11.5%, and the mean peak to peak amplitude changed by 6.8% for the three patients. The consistent image registration generated dense displacement vector fields which represented physiologically reasonable lung motion. In one patient, we observed that the motion patterns are similar in the inferior and posterior parts of the lung, but are different at the superior and anterior parts of the lung. In the other two patients, we observed less variation in lung motion patterns between the two studies. The interfraction motion at end of exhalation is generally small (< 3 mm). Conclusions: The results suggest that lung motion variation may be large for some parts of the lung in some patients. An updated 4D CT study may be needed on a treatment day for some patients. For other patients, one may use the 4D model built on a planning day to estimate lung motion with an external respiratory signal. Quantitative evaluations are under investigation.

Coupled B-snake grids and constrained thin-plate splines for analysis of 2-D tissue deformations from tagged MRI.

Magnetic resonance imaging (MRI) is unique in its ability to noninvasively and selectively alter tissue magnetization and create tagged patterns within a deforming body such as the heart muscle. The resulting patterns define a time-varying curvilinear coordinate system on the tissue, which we track with coupled B-snake grids. B-spline bases provide local control of shape, compact representation, and parametric continuity. Efficient spline warps are proposed which warp an area in the plane such that two embedded snake grids obtained from two tagged frames are brought into registration, interpolating a dense displacement vector field. The reconstructed vector field adheres to the known displacement information at the intersections, forces corresponding snakes to be warped into one another, and for all other points in the plane, where no information is available, a C1 continuous vector field is interpolated. The implementation proposed in this paper improves on our previous variational-based implementation and generalizes warp methods to include biologically relevant contiguous open curves, in addition to standard landmark points. The methods are validated with a cardiac motion simulator, in addition to in-vivo tagging data sets.

Scene segmentation from dense displacement vector fields using randomized Hough transform

Abstract A method which combines two different paradigms for segmenting an image scene by evaluating dense displacement vector fields is presented. First, a rough but robust decomposition of the vector field is achieved by using randomized Hough transform, a technique which is independent from prior knowledge about the actual number of segments in the scene. Subsequently, a merging step fuses those segments most likely belonging to the same object. Finally, a refinement of the segmentation mask is attained by means of a maximum a posteriori (MAP) criterion. To this end the mask is modelled as a Gibbs-Markov random field under the assumption that the scene objects are spatially continuous and only moving slowly between consecutive image frames.

On the estimation of dense displacement vector fields from image sequences (abstract only)

Based on recent experimental as well as theoretical investigations, a generalization of previously published approaches towards the estimation of displacement vector fields is formulated. The calculus of variation allows to transform this approach into a set of two partial differential equations for the two components of the displacement vector field. Some simplifying assumptions facilitate the derivation of an iterative solution approach which can be studied in closed form.