scholarly journals FiberNET: An ensemble deep learning framework for clustering white matter fibers

2017 ◽  
Author(s):  
Vikash Gupta ◽  
Sophia I. Thomopoulos ◽  
Faisal M. Rashid ◽  
Paul M. Thompson
Keyword(s):  
Deep Learning ◽  
White Matter ◽  
Structural Integrity ◽  
Region Of Interest ◽  
Fiber Bundles ◽  
Anatomical Connectivity ◽  
Learning Framework ◽  
Fiber Clustering ◽  
Learning Techniques ◽  
The Brain

AbstractWhite matter tracts are commonly analyzed in studies of micro-structural integrity and anatomical connectivity in the brain. Over the last decade, it has been an open problem as to how best to cluster white matter fibers, extracted from whole-brain tractography, into anatomically meaningful groups. Some existing techniques use region of interest (ROI) based clustering, atlas-based labeling, or unsupervised spectral clustering. ROI-based clustering is popular for analyzing anatomical connectivity among a set of ROIs, but it does not always partition the brain into recognizable fiber bundles. Here we propose an approach using convolutional neural networks (CNNs) to learn shape features of the fiber bundles, which are then exploited to cluster white matter fibers. To achieve such clustering, we first need to re-parameterize the fibers in an intrinsic space. The clustering is performed in induced parameterized coordinates. To our knowledge, this is one of the first approaches for fiber clustering using deep learning techniques. The results show strong accuracy - on a par with or better than other state-of-the-art methods.

scholarly journals Fibernet 2.0: An Automatic Neural Network Based Tool for Clustering White Matter Fibers in the Brain

2017 ◽  
Author(s):  
Vikash Gupta ◽  
Sophia I. Thomopoulos ◽  
Conor K. Corbin ◽  
Faisal Rashid ◽  
Paul M. Thompson
Keyword(s):  
White Matter ◽  
Spectral Clustering ◽  
Developmental Disorders ◽  
Fiber Bundles ◽  
Neural Pathways ◽  
Fiber Tracts ◽  
Learning Techniques ◽  
Segmentation Methods ◽  
White Matter Fiber Tracts ◽  
The Brain

ABSTRACTThe brain’s white matter fiber tracts are impaired in a range of common and devastating conditions, from Alzheimer’s disease to brain trauma, and in developmental disorders such as autism and neurogenetic syndromes. Many studies now examine the connectivity and microstructure of the brain’s neural pathways, spurring the development of algorithms to extract and measure tracts and fiber bundles. Clustering white matter (WM) fibers, from whole-brain tractography, into anatomically meaningful bundles is still a challenging problem. Existing tract segmentation methods use atlases or regions of interest (ROI) or unsupervised spectral clustering. Even so, atlas-based segmentation does not always partition the brain into a set of recognizable fiber bundles. Deep learning techniques can be applied to automatically segment and cluster white matter fibers. Here we propose a robust approach using convolutional neural networks (CNNs) to learn shape features of the fiber bundles, which we then exploit to cluster WM fibers into bundles. In a range of tests across diverse fiber bundles, we illustrate the accuracy of our method, and its ability to suppress false positive fibers.

scholarly journals Mol-BERT: An Effective Molecular Representation with BERT for Molecular Property Prediction

2021 ◽  
Vol 2021 ◽  
pp. 1-7
Author(s):  
Juncai Li ◽  
Xiaofei Jiang
Keyword(s):  
Deep Learning ◽  
Language Processing ◽  
Large Scale ◽  
Molecular Data ◽  
Molecular Property ◽  
Property Prediction ◽  
Learning Framework ◽  
Learning Techniques ◽  
Potential Benefits ◽  
Current Sequence

Molecular property prediction is an essential task in drug discovery. Most computational approaches with deep learning techniques either focus on designing novel molecular representation or combining with some advanced models together. However, researchers pay fewer attention to the potential benefits in massive unlabeled molecular data (e.g., ZINC). This task becomes increasingly challenging owing to the limitation of the scale of labeled data. Motivated by the recent advancements of pretrained models in natural language processing, the drug molecule can be naturally viewed as language to some extent. In this paper, we investigate how to develop the pretrained model BERT to extract useful molecular substructure information for molecular property prediction. We present a novel end-to-end deep learning framework, named Mol-BERT, that combines an effective molecular representation with pretrained BERT model tailored for molecular property prediction. Specifically, a large-scale prediction BERT model is pretrained to generate the embedding of molecular substructures, by using four million unlabeled drug SMILES (i.e., ZINC 15 and ChEMBL 27). Then, the pretrained BERT model can be fine-tuned on various molecular property prediction tasks. To examine the performance of our proposed Mol-BERT, we conduct several experiments on 4 widely used molecular datasets. In comparison to the traditional and state-of-the-art baselines, the results illustrate that our proposed Mol-BERT can outperform the current sequence-based methods and achieve at least 2% improvement on ROC-AUC score on Tox21, SIDER, and ClinTox dataset.

scholarly journals Dental Images’ Segmentation Using Threshold Connected Component Analysis

2021 ◽  
Vol 2021 ◽  
pp. 1-9
Author(s):  
Vincent Majanga ◽  
Serestina Viriri
Keyword(s):  
Deep Learning ◽  
Region Of Interest ◽  
Component Analysis ◽  
Connected Components ◽  
Connected Component ◽  
Low Contrast ◽  
Connected Component Analysis ◽  
Feature Representations ◽  
Learning Techniques ◽  
Performance Results

Recent advances in medical imaging analysis, especially the use of deep learning, are helping to identify, detect, classify, and quantify patterns in radiographs. At the center of these advances is the ability to explore hierarchical feature representations learned from data. Deep learning is invaluably becoming the most sought out technique, leading to enhanced performance in analysis of medical applications and systems. Deep learning techniques have achieved great performance results in dental image segmentation. Segmentation of dental radiographs is a crucial step that helps the dentist to diagnose dental caries. The performance of these deep networks is however restrained by various challenging features of dental carious lesions. Segmentation of dental images becomes difficult due to a vast variety in topologies, intricacies of medical structures, and poor image qualities caused by conditions such as low contrast, noise, irregular, and fuzzy edges borders, which result in unsuccessful segmentation. The dental segmentation method used is based on thresholding and connected component analysis. Images are preprocessed using the Gaussian blur filter to remove noise and corrupted pixels. Images are then enhanced using erosion and dilation morphology operations. Finally, segmentation is done through thresholding, and connected components are identified to extract the Region of Interest (ROI) of the teeth. The method was evaluated on an augmented dataset of 11,114 dental images. It was trained with 10 090 training set images and tested on 1024 testing set images. The proposed method gave results of 93 % for both precision and recall values, respectively.

scholarly journals Early white matter development is abnormal in tuberous sclerosis complex patients who develop autism spectrum disorder

2019 ◽  
Vol 11 (1) ◽  
Author(s):  
Anna K. Prohl ◽  
◽  
Benoit Scherrer ◽  
Xavier Tomas-Fernandez ◽  
Peter E. Davis ◽  
...  
Keyword(s):  
Autism Spectrum Disorder ◽  
White Matter ◽  
Tuberous Sclerosis ◽  
Tuberous Sclerosis Complex ◽  
Neural Circuits ◽  
Diffusion Tensor ◽  
Autism Spectrum ◽  
Spectrum Disorder ◽  
Fiber Bundles ◽  
The Brain

Abstract Background Autism spectrum disorder (ASD) is prevalent in tuberous sclerosis complex (TSC), occurring in approximately 50% of patients, and is hypothesized to be caused by disruption of neural circuits early in life. Tubers, or benign hamartomas distributed stochastically throughout the brain, are the most conspicuous of TSC neuropathology, but have not been consistently associated with ASD. Widespread neuropathology of the white matter, including deficits in myelination, neuronal migration, and axon formation, exist and may underlie ASD in TSC. We sought to identify the neural circuits associated with ASD in TSC by identifying white matter microstructural deficits in a prospectively recruited, longitudinally studied cohort of TSC infants. Methods TSC infants were recruited within their first year of life and longitudinally imaged at time of recruitment, 12 months of age, and at 24 months of age. Autism was diagnosed at 24 months of age with the ADOS-2. There were 108 subjects (62 TSC-ASD, 55% male; 46 TSC+ASD, 52% male) with at least one MRI and a 24-month ADOS, for a total of 187 MRI scans analyzed (109 TSC-ASD; 78 TSC+ASD). Diffusion tensor imaging properties of multiple white matter fiber bundles were sampled using a region of interest approach. Linear mixed effects modeling was performed to test the hypothesis that infants who develop ASD exhibit poor white matter microstructural integrity over the first 2 years of life compared to those who do not develop ASD. Results Subjects with TSC and ASD exhibited reduced fractional anisotropy in 9 of 17 white matter regions, sampled from the arcuate fasciculus, cingulum, corpus callosum, anterior limbs of the internal capsule, and the sagittal stratum, over the first 2 years of life compared to TSC subjects without ASD. Mean diffusivity trajectories did not differ between groups. Conclusions Underconnectivity across multiple white matter fiber bundles develops over the first 2 years of life in subjects with TSC and ASD. Future studies examining brain-behavior relationships are needed to determine how variation in the brain structure is associated with ASD symptoms.

A Study on Brain Tumor Detection and Segmentation Using Deep Learning Techniques

2020 ◽  
Vol 17 (4) ◽  
pp. 1925-1930
Author(s):  
Ambeshwar Kumar ◽  
R. Manikandan ◽  
Robbi Rahim
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Deep Learning ◽  
Brain Tumor ◽  
Tumor Detection ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
Learning Techniques ◽  
Convolution Algorithm ◽  
The Brain

It’s a new era technology in the field of medical engineering giving awareness about the various healthcare features. Deep learning is a part of machine learning, it is capable of handling high dimensional data and is efficient in concentrating on the right features. Tumor is an unbelievably complex disease: a multifaceted cell has more than hundred billion cells; each cell acquires mutation exclusively. Detection of tumor particles in experiment is easily done by MRI or CT. Brain tumors can also be detected by MRI, however, deep learning techniques give a better approach to segment the brain tumor images. Deep Learning models are imprecisely encouraged by information handling and communication designs in biological nervous system. Classification plays an significant role in brain tumor detection. Neural network is creating a well-organized rule for classification. To accomplish medical image data, neural network is trained to use the Convolution algorithm. Multilayer perceptron is intended for identification of a image. In this study article, the brain images are categorized into two types: normal and abnormal. This article emphasize the importance of classification and feature selection approach for predicting the brain tumor. This classification is done by machine learning techniques like Artificial Neural Networks, Support Vector Machine and Deep Neural Network. It could be noted that more than one technique can be applied for the segmentation of tumor. The several samples of brain tumor images are classified using deep learning algorithms, convolution neural network and multi-layer perceptron.

scholarly journals Minimal Linear Networks for Magnetic Resonance Image Reconstruction

2019 ◽  
Vol 9 (1) ◽  
Author(s):  
Gilad Liberman ◽  
Benedikt A. Poser
Keyword(s):  
Deep Learning ◽  
Magnetic Resonance ◽  
Image Reconstruction ◽  
Simulated Data ◽  
Linear Networks ◽  
Learning Framework ◽  
Scan Time ◽  
Learning Techniques ◽  
Ill Posed ◽  
Magnetic Resonance Imaging Mri

AbstractModern sequences for Magnetic Resonance Imaging (MRI) trade off scan time with computational challenges, resulting in ill-posed inverse problems and the requirement to account for more elaborated signal models. Various deep learning techniques have shown potential for image reconstruction from reduced data, outperforming compressed sensing, dictionary learning and other advanced techniques based on regularization, by characterization of the image manifold. In this work we suggest a framework for reducing a “neural” network to the bare minimum required by the MR physics, reducing the network depth and removing all non-linearities. The networks performed well both on benchmark simulated data and on arterial spin labeling perfusion imaging, showing clear images while preserving sensitivity to the minute signal changes. The results indicate that the deep learning framework plays a major role in MR image reconstruction, and suggest a concrete approach for probing into the contribution of additional elements.

Fractional anisotropy and mean diffusivity parameters of the brain white matter tracts in preterm infants: reproducibility of region-of-interest measurements

2011 ◽  
Vol 42 (2) ◽  
pp. 175-182 ◽  
Author(s):  
Virva K. Lepomäki ◽  
◽  
Teemu P. Paavilainen ◽  
Saija A. M. Hurme ◽  
Markku E. Komu ◽  
...  
Keyword(s):  
White Matter ◽  
Fractional Anisotropy ◽  
Preterm Infants ◽  
Mean Diffusivity ◽  
Region Of Interest ◽  
White Matter Tracts ◽  
Brain White Matter ◽  
The Brain

Robust fiber clustering of cerebral fiber bundles in white matter

2014 ◽  
Author(s):  
Xufeng Yao ◽  
Yongxiong Wang ◽  
Songlin Zhuang
Keyword(s):  
White Matter ◽  
Fiber Bundles ◽  
Fiber Clustering

scholarly journals Deep Learning Based Classification of Radar Spectral Maps

2021 ◽  
pp. 99-104
Author(s):  
Tong Lin ◽  
◽  
Xin Chen ◽  
Xiao Tang ◽  
Ling He ◽  
...  
Keyword(s):  
Deep Learning ◽  
Classification Accuracy ◽  
Region Of Interest ◽  
Radar Target ◽  
Automatic Annotation ◽  
Deep Convolutional Neural Networks ◽  
Learning Techniques ◽  
Using Data ◽  
Map Data

This paper discusses the use of deep convolutional neural networks for radar target classification. In this paper, three parts of the work are carried out: firstly, effective data enhancement methods are used to augment the dataset and address unbalanced datasets. Second, using deep learning techniques, we explore an effective framework for classifying and identifying targets based on radar spectral map data. By using data enhancement and the framework, we achieved an overall classification accuracy of 0.946. In the end, we researched the automatic annotation of image ROI (region of interest). By adjusting the model, we obtained a 93% accuracy in automatic labeling and classification of targets for both car and cyclist categories.

scholarly journals Parallel Deep Learning Framework for Video Surveillance System

2021 ◽  
Author(s):  
Mayuri Karvande ◽  
Apoorv Katkar ◽  
Nikhil Koli ◽  
Amit Joshi ◽  
Suraj Sawant
Keyword(s):  
Deep Learning ◽  
Video Surveillance ◽  
System Performance ◽  
System Stability ◽  
Camera System ◽  
Public Places ◽  
Dynamic Selection ◽  
Learning Framework ◽  
Learning Techniques ◽  
The Right

In today’s world, the security of every individual has become an important aspect. There is a need for constant monitoring in public places. A Manual operating camera system is an unreliable and very basic and poor method for this purpose. Intelligent Video Surveillance is an approach where multiple CCTVs constantly record the scenes and proper algorithms are deployed in order to detect and monitor activities. Deep Learning frameworks and algorithms like Kera’s, YOLO, Convolutional Neural Networks or backbones for image detection like VGG16, Mobile net, Resnet101 have been used for human and weapon detection. The paper focuses on deep learning techniques and threading to collectively develop a Parallel Deep Learning Framework for Video Surveillance that aims at striking the right balance between accuracy and system performance or stability. Threading is used in terms of implementation of a uniquely proposed Dynamic Selection Algorithm that uses two backbones for object detection and switches between them based on the queue status for achieving system stability. A uniquely designed logistic regression filter is also implemented that boosts the system performance.

Sign in / Sign up

Export Citation Format

Share Document