Robust brain ROI segmentation by deformation regression and deformable shape model.

This paper proposes a novel method of using regression-guided deformable models for brain regions of interest (ROIs) segmentation. Different from conventional deformable segmentation, which often deforms shape model locally and thus sensitive to initialization, we propose to learn a regressor to explicitly guide the shape deformation, thus eventually improves the performance of ROI segmentation. The regressor is learned via two steps, (1) a joint classification and regression random forest (CRRF) and (2) an auto-context model. The CRRF predicts each voxel's deformation to the nearest point on the ROI boundary as well as each voxel's class label (e.g., ROI versus background). The auto-context model further refines all voxel's deformations (i.e., deformation field) and class labels (i.e., label maps) by considering the neighboring structures. Compared to the conventional random forest regressor, the proposed regressor provides more accurate deformation field estimation and thus more robust in guiding deformation of the shape model. Validated in segmentation of 14 midbrain ROIs from the IXI dataset, our method outperforms the state-of-art multi-atlas label fusion and classification methods, and also significantly reduces the computation cost.

Validating a new methodology for optical probe design and image registration in fNIRS studies

Objective: The Vestibulocerebellar Tract (VCT) connects the vestibular nuclear complex mainly with the uvula-nodulus of the cerebellum. Many animal studies have revealed the anatomy of the VCT, but very little is known about the VCT in the human brain. In this study, we attempted to identify the VCT in the normal human brain using diffusion tensor tractography. Methods: Twenty healthy volunteers were recruited for this study. Diffusion tensor imaging data was acquired using 1.5-T, and the VCT was obtained using FMRIB software. The VCT was determined by selection of fibers passing through two Regions of Interest (ROIs): seed ROI - vestibular nuclei areas, target ROI - the uvula of the cerebellum. Results: VCTs that originated from the vestibular nuclei passed posteriorly via the inferior cerebellar peduncle, then terminated at the uvula of cerebellum. No significant differences were observed in terms of fractional anisotropy, mean diffusivity, and tract volume of the VCT according to the side of the hemisphere (p > 0.05). Conclusion: We identified the VCT in the human brain using diffusion tensor tractography and believe that the methodology and results reported here will be helpful to researchers and clinicians in this field. Keywords: Vestibulocerebellar tract, diffusion tensor tractography, uvula, cerebellum, vestibular nuclear complex, brain.

This paper presents a new deformable modeling strategy aimed at integrating shape and appearance in a unified space. If we think traditional deformable models as active contours or evolving curve fronts, the new deformable shape and appearance models we propose are deforming disks or volumes. Each model has not only boundary shape but also interior appearance. The model shape is implicitly embedded in a higher dimensional space of distance transforms, thus represented by a distance map image. In this way, both shape and appearance of the model are defined in the pixel space. A common deformation scheme, the Free Form Deformations (FFD), parameterizes warping deformations of the volumetric space in which the model is embedded in, hence deforming both model boundary and interior simultaneously.

Purpose: The ascending reticular activating system (ARAS) is responsible for regulation of consciousness. In this study, using diffusion tensor imaging (DTI), we attempted to reconstruct the thalamocortical projections between the intralaminar thalamic nuclei and the frontoparietal cortex in normal subjects. Methods: DTI data were acquired in 24 healthy subjects and eight kinds of thalamocortical projections were reconstructed: the seed re gion of interest (ROI) - the intralaminar thalamic nuclei and the eight target ROIs - the medial prefrontal cortex, dorsolateral prefrontal cortex, ventrolateral prefrontal cortex, orbitofrontal cortex, premotor cortex, primary motor cortex, primary somatosensory cortex, and posterior parietal cortex. Results: The eight thalamocortical projections were reconstructed in each hemisphere and the pathways were visualized: projections to the prefrontal cortex ascended through the anterior limb and genu of the internal capsule and anterior corona radiata. Projections to the premotor cortex passed through the genu and posterior limb of the internal capsule and middle corona radiata; in contrast, projections to the primary motor cortex, primary somatosensory cortex, and posterior parietal cortex ascended through the posterior limb of the in ternal capsule. No significant difference in fractional anisotropy, mean diffusivity, and fiber volume of all reconstructed thalamocortical projections was observed between the right and left hemispheres (p>0.05). Conclusion: We reconstructed the thalamocortical projections between the intralaminar thalamic nuclei and the frontoparietal cortex in normal subjects. We believe that our findings would be useful to clinicians involved in the care of patients with impaired consciousness and for researchers in studies of the ARAS.

Traumatic axonal injury of the medial lemniscus pathway in a patient with traumatic brain injury: validation by diffusion tensor tractography.

Segmentation of left ventricular (LV) endocardium from 3D echocardiography is important for clinical diagnosis because it not only can provide some clinical indices (e.g. ventricular volume and ejection fraction) but also can be used for the analysis of anatomic structure of ventricle. In this work, we proposed a new full-automatic method, combining the deep learning and deformable model, for the segmentation of LV endocardium. We trained convolutional neural networks to generate a binary cuboid to locate the region of interest (ROI). And then, using ROI as the input, we trained stacked autoencoder to infer the LV initial shape. At last, we adopted snake model initiated by inferred shape to segment the LV endocardium. In the experiments, we used 3DE data, from CETUS challenge 2014 for training and testing by segmentation accuracy and clinical indices. The results demonstrated the proposed method is accuracy and efficiency respect to expert's measurements.

In this study, we propose a semantic grasping method that accounts for post-grasp stability without requiring a preexisting 3D shape of the object. Among multiple candidate grasp points, the method selects the one that minimizes the moment at the grasp point with respect to the object’s centroid. This paper proposes a semantic grasping method that explicitly considers post-grasp stability without requiring an instance-specific 3D shape model. The method estimates a grasp pose that minimizes moment at the grasp point with respect to the object’s centroid, thereby maintaining a stable orientation after lifting. In experiments, we evaluated post-grasp stability against conventional methods; the proposed method maintained a stable state in 18 out of 25 trials, outperforming the conventional method (4 out of 25).

Single-pixel camera is developed to mitigate the constraints faced by the conventional cameras especially in invisible wavelengths and low light conditions. Nyquist–Shannon theorem requires as many measurements as the image pixels to reconstruct images flawlessly. In practice, obtaining more measurements increases the cost and acquisition time, which are the major drawbacks of single-pixel imaging (SPI). Therefore, compressive sensing was proposed to enable image reconstruction with fewer measurements. We present a design of sensing patterns to obtain image information by utilizing spatially variant resolution (SVR) technique in SPI. The proposed method reduces the measurements by prioritizing the resolution in the region of interest (ROI). It successfully achieves the programmable imaging concept where multiresolution adaptively optimizes the balance between the image quality and the measurements number. Results show that SVR images can be reconstructed from significantly fewer measurements yet able to achieve better image quality than uniform resolution images. In addition, the SVR images can be further enhanced by integrating the dynamic supersampling technique. Consequently, the concerns of image quality, long acquisition, and processing time can be addressed. The proposed method potentially benefits imaging applications where the target ROI is prioritized over the background and most importantly it requires fewer measurements.

Central vestibular disorder is common after middle cerebral artery (MCA) territory infarction. The MCA supplies blood to the parieto-insular vestibular cortex (PIVC), a core region of central vestibular symptoms. We report on patients that sustained injuries of the core vestibular pathway to the PIVC with central vestibular disorder following MCA territory infarction, demonstrated on diffusion tensor imaging (DTI). Nineteen patients with MCA territory infarction and 12 control subjects were recruited. To reconstruct the core vestibular pathway to the PIVC, we defined seed region of interest (ROI) as vestibular nuclei of pons and target ROI as the PIVC. Fractional anisotropy (FA), mean diffusivity, and tract volume were measured. In the affected hemisphere, FA value of the core vestibular pathway to the PIVC revealed significant difference between all patient groups and the control group (P < .05). In contrast, patients with symptoms of ataxia only revealed significant decrement of tract volume compared with the control group (P < .05). Additionally, subgroup B revealed significant decrement of tract volume compared with that of subgroup A and the control group (P < .05). In the unaffected hemisphere, there was no significant difference in all DTI parameters between all patient groups and the control group (P < .05). Injury to the core vestibular pathway to the PIVC was demonstrated in patients that revealed typical central vestibular disorder following MCA territory infarction. Analysis of the core vestibular pathway to the PIVC using DTI would be beneficial in clinical evaluation and management of patients with MCA territory infarction.

Lateral medullary syndrome following injury of the vestibular pathway to the core vestibular cortex: Diffusion tensor imaging study.

Pyrosequencing is a "sequencing by synthesis" technique which can be used to quantify DNA methylation at specific CpG sites within the target region of interest. Biotin labelled polymerase chain reaction (PCR) products form the template for base-pair nucleotide incorporation causing a light emitting cascade reaction resulting in the formation of a pyrogram and the calculation of the percentage methylation for each site. Prior to pyrosequencing, it is essential to bisulphite-convert the DNA sample and then perform locus-specific PCR for the region of interest. One of the PCR primers needs to be biotinylated and a separate sequencing primer is required for the pyrosequencing itself.

In functional MRI, magnetic field inhomogeneities due to air-tissue susceptibility differences may lead to severe signal dropouts and geometric distortions in echo-planar images. Therefore, the inhomogeneities in the field are routinely minimized by shimming prior to imaging. However in fMRI, the Blood Oxygen Level Dependent (BOLD) effect is the measure of interest, so the BOLD sensitivity (BS) should be optimized rather than the magnetic field homogeneity. The analytical expression for an estimate of the BOLD sensitivity has been recently developed, allowing for the computation of BOLD sensitivity maps from echo-planar images and field maps. This report describes a novel shimming procedure that optimizes the local BOLD sensitivity over a region of interest. The method is applied in vivo and compared to a standard global shimming procedure. A breath-holding experiment was carried out and demonstrated that the BS-based shimming significantly improved the detection of activation in a target region of interest, the medial orbitofrontal cortex.

A 3D caricature is an exaggerated 3D depiction of a human face. The goal of this paper is to model the variations of 3D caricatures in a compact parameter space so that we can provide a useful data-driven toolkit for handling 3D caricature deformations. To achieve the goal, we propose an MLP-based framework for building a deformable surface model, which takes a latent code and produces a 3D surface. In the framework, a SIREN MLP models a function that takes a 3D position on a fixed template surface and returns a 3D displacement vector for the input position. We create variations of 3D surfaces by learning a hypernetwork that takes a latent code and produces the parameters of the MLP. Once learned, our deformable model provides a nice editing space for 3D caricatures, supporting label-based semantic editing and point-handle-based deformation, both of which produce highly exaggerated and natural 3D caricature shapes. We also demonstrate other applications of our deformable model, such as automatic 3D caricature creation.

Using the knowledge of information technology and physical mechanics to simulate the evolution of plant cell shapes is helpful to explore the mechanism of plant tissue and organ formation. However, the traditional numerical modeling based on curves and surfaces is difficult to accurately locate a geometric fragment of cell shape, resulting in the inability to analyze the stress and strain of the fragment freely. In this study, a plant cell shape is firstly decomposed into a body, faces, edges, and vertexes from top to bottom using the n-dimensional generalized mapping theory. Then, the 1-, 2-and 3-dimensional topologies of the cell shape are constructed by combining these topologies from bottom to top using the “half-edge” structure of 0-dimensional topology, and the corresponding retrieval algorithms are designed. Finally, the program is achieved by the C++ programming language and the OpenGL graphics library. The experimental results show that users could not only retrieve the geometric units of cell shape freely, but also highlight the query results in the program through the query interface of this program. The results of this study can provide an automatic topological retrieval of a cell shape for the elastic mechanical modeling of cell shape deformation.