An official website of the United States government
Segmentation of multiple surfaces in optical coherence tomography (OCT) images is a challenging problem, further complicated by the frequent presence of weak boundaries, varying layer thicknesses, and mutual influence between adjacent surfaces. The traditional graph-based optimal surface segmentation method has proven its effectiveness with its ability to capture various surface priors in a uniform graph model. However, its efficacy heavily relies on handcrafted features that are used to define the surface cost for the “goodness” of a surface. Recently, deep learning (DL) is emerging as a powerful tool for medical image segmentation thanks to its superior feature learning capability. Unfortunately, due to the scarcity of training data in medical imaging, it is nontrivial for DL networks toimplicitlylearn the global structure of the target surfaces, including surface interactions. This study proposes to parameterize the surface cost functions in the graph model and leverage DL to learn those parameters. The multiple optimal surfaces are then simultaneously detected by minimizing the total surface cost whileexplicitlyenforcing the mutual surface interaction constraints. The optimization problem is solved by the primal-dual interior-point method (IPM), which can be implemented by a layer of neural networks, enabling efficient end-to-end training of the whole network. Experiments on spectral-domain optical coherence tomography (SD-OCT) retinal layer segmentation demonstrated promising segmentation results with sub-pixel accuracy.
more »
« less
Learning Cortical Parcellations Using Graph Neural Networks
Deep learning has been applied to magnetic resonance imaging (MRI) for a variety of purposes, ranging from the acceleration of image acquisition and image denoising to tissue segmentation and disease diagnosis. Convolutional neural networks have been particularly useful for analyzing MRI data due to the regularly sampled spatial and temporal nature of the data. However, advances in the field of brain imaging have led to network- and surface-based analyses that are often better represented in the graph domain. In this analysis, we propose a general purpose cortical segmentation method that, given resting-state connectivity features readily computed during conventional MRI pre-processing and a set of corresponding training labels, can generate cortical parcellations for new MRI data. We applied recent advances in the field of graph neural networks to the problem of cortical surface segmentation, using resting-state connectivity to learn discrete maps of the human neocortex. We found that graph neural networks accurately learn low-dimensional representations of functional brain connectivity that can be naturally extended to map the cortices of new datasets. After optimizing over algorithm type, network architecture, and training features, our approach yielded mean classification accuracies of 79.91% relative to a previously published parcellation. We describe how some hyperparameter choices including training and testing data duration, network architecture, and algorithm choice affect model performance.
more »
« less
- Award ID(s):
- 1734430
- PAR ID:
- 10397889
- Date Published:
- Journal Name:
- Frontiers in Neuroscience
- Volume:
- 15
- ISSN:
- 1662-453X
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
A central goal in neuroscience is to understand how dynamic networks of neural activity produce effective representations of the world. Advances in the theory of graph measures raise the possibility of elucidating network topologies central to the construction of these representations. We leverage a result from the description of lollipop graphs to identify an iconic network topology in functional magnetic resonance imaging data and characterize changes to those networks during task performance and in populations diagnosed with psychiatric disorders. During task performance, we find that task-relevant subnetworks change topology, becoming more integrated by increasing connectivity throughout cortex. Analysis of resting-state connectivity in clinical populations shows a similar pattern of subnetwork topology changes; resting-scans becoming less default-like with more integrated sensory paths. The study of brain network topologies and their relationship to cognitive models of information processing raises new opportunities for understanding brain function and its disorders.more » « less
-
Iron overload, a complication of repeated blood transfusions, can cause tissue damage and organ failure. The body has no regulatory mechanism to excrete excess iron, so iron overload must be closely monitored to guide therapy and measure treatment response. The concentration of iron in the liver is a reliable marker for total body iron content and is now measured noninvasively with magnetic resonance imaging (MRI). MRI produces a diagnostic image by measuring the signals emitted from the body in the presence of a constant magnetic field and radiofrequency pulses. At each pixel, the signal decay constant, T2*, can be calculated, providing insight about the structure of each tissue. Liver iron content can be quantified based on this T2* value because signal decay accelerates with increasing iron concentration. We developed a method to automatically segment the liver from the MRI image to accurately calculate iron content. Our current algorithm utilizes the active contour model for image segmentation, which iteratively evolves a curve until it reaches an edge or a boundary. We applied this algorithm to each MRI image in addition to a map of pixelwise T2* values, combining basic image processing with imaging physics. One of the limitations of this segmentation model is how it handles noise in the MRI data. Recent advancements in deep learning have enabled researchers to utilize convolutional neural networks to denoise and reconstruct images. We used the Trainable Nonlinear Reaction Diffusion network architecture to denoise the MRI images, allowing for smoother segmentation while preserving fine details.more » « less
-
Advances in graph signal processing for network neuroscience offer a unique pathway to integrate brain structure and function, with the goal of revealing some of the brain's organizing principles at the system level. In this direction, we develop a supervised graph representation learning framework to model the relationship between brain structural connectivity (SC) and functional connectivity (FC) via a graph encoder-decoder system. Specifically, we propose a Siamese network architecture equipped with graph convolutional encoders to learn graph (i.e., subject)-level embeddings that preserve application-dependent similarity measures between brain networks. This way, we effectively increase the number of training samples and bring in the flexibility to incorporate additional prior information via the prescribed target graph-level distance. While information on the brain structure-function coupling is implicitly distilled via reconstruction of brain FC from SC, our model also manages to learn representations that preserve the similarity between input graphs. The superior discriminative power of the learnt representations is demonstrated in downstream tasks including subject classification and visualization. All in all, this work advocates the prospect of leveraging learnt graph-level, similarity-preserving embeddings for brain network analysis, by bringing to bear standard tools of metric data analysis.more » « less
-
This work presents a novel deep learning architecture called BNU-Net for the purpose of cardiac segmentation based on short-axis MRI images. Its name is derived from the Batch Normalized (BN) U-Net architecture for medical image segmentation. New generations of deep neural networks (NN) are called convolutional NN (CNN). CNNs like U-Net have been widely used for image classification tasks. CNNs are supervised training models which are trained to learn hierarchies of features automatically and robustly perform classification. Our architecture consists of an encoding path for feature extraction and a decoding path that enables precise localization. We compare this approach with a parallel approach named U-Net. Both BNU-Net and U-Net are cardiac segmentation approaches: while BNU-Net employs batch normalization to the results of each convolutional layer and applies an exponential linear unit (ELU) approach that operates as activation function, U-Net does not apply batch normalization and is based on Rectified Linear Units (ReLU). The presented work (i) facilitates various image preprocessing techniques, which includes affine transformations and elastic deformations, and (ii) segments the preprocessed images using the new deep learning architecture. We evaluate our approach on a dataset containing 805 MRI images from 45 patients. The experimental results reveal that our approach accomplishes comparable or better performance than other state-of-the-art approaches in terms of the Dice coefficient and the average perpendicular distance. Index Terms—Magnetic Resonance Imaging; Batch Normalization; Exponential Linear Unitsmore » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3389/fnins.2021.797500
