A Combined Deep-Learning and Lattice Boltzmann Model for Segmentation of the Hippocampus in MRI

Abstract

<jats:p>Segmentation of the hippocampus (HC) in magnetic resonance imaging (MRI) is an essential step for diagnosis and monitoring of several clinical situations such as Alzheimer’s disease (AD), schizophrenia and epilepsy. Automatic segmentation of HC structures is challenging due to their small volume, complex shape, low contrast and discontinuous boundaries. The active contour model (ACM) with a statistical shape prior is robust. However, it is difficult to build a shape prior that is general enough to cover all possible shapes of the HC and that suffers the problems of complicated registration of the shape prior and the target object and of low efficiency. In this paper, we propose a semi-automatic model that combines a deep belief network (DBN) and the lattice Boltzmann (LB) method for the segmentation of HC. The training process of DBN consists of unsupervised bottom-up training and supervised training of a top restricted Boltzmann machine (RBM). Given an input image, the trained DBN is utilized to infer the patient-specific shape prior of the HC. The specific shape prior is not only used to determine the initial contour, but is also introduced into the LB model as part of the external force to refine the segmentation. We used a subset of OASIS-1 as the training set and the preliminary release of EADC-ADNI as the testing set. The segmentation results of our method have good correlation and consistency with the manual segmentation results.</jats:p>

Hippocampal Segmentation in Brain MRI Images Using Machine Learning Methods: A Survey

“…The slightly larger 2D mask is merged with the original MRI image to obtain the pre‐segmented hippocampus, and then the incorrect labels are corrected by replacing and thinning the network. Liu et al [ 87 ] proposed a semiautomatic model that combines a Deep belief network (DBN) and the Lattice Boltzmann (LB) method to segment the hippocampus. Given the input image, the trained DBN is used to infer the patient‐specific shape prior for the hippocampus.…”

“…There are many datasets, such as HarP [ 105 ] , ABIDE [ 90 ] and IBSR [ 106 ] , that can provide a small number of samples with manual segmentation results. Some datasets, such as OASIS [ 87 ] , provide a large number of images without manual segmentation; some datasets provide brain MRI images of specific populations, such as patients with Alzheimer's disease, epilepsy patients, and babies; and other datasets, such as MMMRR [ 107 ] , provide multimodal images of the brain. Finally, the parent project of the LONI dataset provides brain MRI images of multiple species.…”

Deep Learning-Based Intelligent Robot in Sentencing

“…It can learn data without supervision. Figure 13 is a schematic diagram of the network structure of the restricted Boltzmann deep learning model ( Liu and Yan, 2020 ). Figure 13 shows that the restricted Boltzmann deep learning model is composed of two layers of neural networks: the visible layer and the hidden layer.…”

Research on Pathogenic Hippocampal Voxel Detection in Alzheimer's Disease Using Clustering Genetic Random Forest

“…Tsao et al used a convex fused sparse group lasso method and multivariate tensor-based morphometry method to predict the AD features ( 17 ). Liu et al introduced a fusion method using the deep belief network method and the lattice Boltzmann method to segment the MRI image, and the correlation and consistency were compared with manual segmentation methods ( 18 ). Using the SVM, random forest, logistic regression, and K-Nearest Neighbors, Uysal et al analyzed the MRI images to distinguish stages of AD ( 19 ).…”