3D convolutional neural network for feature extraction and classification of fMRI volumes

Vu, Hanh; Kim, Hyun-Chul; Lee, Jong‐Hwan

doi:10.1109/prni.2018.8423964

Cited by 9 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To study whether a supervised classifier could recover the experimenter labels, we first conducted supervised classification of the BOLD data to examine classifier performance when using the folk emotion labels. Specifically, we used a 3D Convolutional Neural Network 54 with sixfold cross validation, iterating across all combinations of training on each participant's data from five runs and testing on a left-out run (see Supplementary Materials for classifier details). Mean within-participant accuracy was calculated by averaging the accuracy across all crossvalidation folds.…”

Section: Resultsmentioning

confidence: 99%

Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience

Azari

Westlin

Satpute

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Machine learning methods provide powerful tools to map physical measurements to scientific categories. But are such methods suitable for discovering the ground truth about psychological categories? We use the science of emotion as a test case to explore this question. In studies of emotion, researchers use supervised classifiers, guided by emotion labels, to attempt to discover biomarkers in the brain or body for the corresponding emotion categories. This practice relies on the assumption that the labels refer to objective categories that can be discovered. Here, we critically examine this approach across three distinct datasets collected during emotional episodes—measuring the human brain, body, and subjective experience—and compare supervised classification solutions with those from unsupervised clustering in which no labels are assigned to the data. We conclude with a set of recommendations to guide researchers towards meaningful, data-driven discoveries in the science of emotion and beyond.

show abstract

Section: Resultsmentioning

confidence: 99%

Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience

Azari

Westlin

Satpute

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…A revised and optimized 3D-CNN model which is formed by Vu et al [13] is used in this study. LeNet-5 [12] CNN network has been used for 2D image classification and Vu et al extended this network to 3D fMRI volume classification.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…Three-dimensional CNN (3D-CNN) is mainly used for 3D images or videos, the kernel slides in 3 dimensions. LeNet-5 [12] CNN network has been used for 2D image classification, and Vu et al [13] extended this network to 3D fMRI volume classification of four sensorimotor tasks. They showed that threedimensional feature maps extracted from fMRI volumes represent brain signals better.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, showing that CNN can learn helpful information from raw data without any extra steps. Also, we introduce the use of the 3D-CNN model for sensorimotor classification proposed by Vu et al's [13] with hyperparameter optimization. Finally, the network performance is tested with the publicly available ADHD datasets and compared with some state-of-the-art models [14][15][16][17] in the literature.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

3D CNN Based Automatic Diagnosis of ADHD Using fMRI Volumes

TAŞPINAR>

Özkurt

2023

Deu Muhendislik Fakultesi Fen Ve Muhendislik

View full text Add to dashboard Cite

Attention deficit hyperactivity disorder (ADHD) is one of the most common mental health disorders and it is threatening especially to the academic performance of children. Its neurobiological diagnosis is essential for clinicians to treat ADHD patients properly. Along with machine learning algorithms, and neuroimaging technologies, especially functional magnetic resonance imaging is increasingly used as biomarker in attention deficit hyperactivity disorder. Also, machine learning methods have been becoming popular at last times. This study presents an optimized 3-dimensional convolutional neural network to classify functional magnetic resonance imaging volumes into two classes to assist experts in diagnosing ADHD. To demonstrate the importance of extracting 3D relationships of data, the method has been tested on ADHD-200 public datasets and its performance on the hold-out testing datasets has been evaluated. Then the network performance has been compared with several recent ADHD detection convolutional neural networks in the literature. It has been observed that the proposed network has a promising performance.

show abstract

“…The ConvBlocks are parameterized by the number of filters inside the convolutional layers. Using those ConvBlocks, the architecture forms a typical encoder decoder architecture in a U-shape, similar to a 3D U-Net [46]. Note that between each ConvBlock in the encoder / decoder, we downsample / upsample the volume by a factor of two and double / halve the number of filters for the next ConvBlock respectively.…”

Section: Neural Network Architecturementioning

confidence: 99%

Deep Volumetric Ambient Occlusion

Engel¹,

Ropinski

2021

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

Full Render with AO AO only Without AO Ground Truth AO Fig. 1: Volume rendering with volumetric ambient occlusion achieved through Deep Volumetric Ambient Occlusion (DVAO). DVAO uses a 3D convolutional encoder-decoder architecture, to predict ambient occlusion volumes for a given combination of volume data and transfer function. While we introduce and compare several representation and injection strategies for capturing the transfer function information, the shown images result from preclassified injection based on an implicit representation.

show abstract

3D convolutional neural network for feature extraction and classification of fMRI volumes

Cited by 9 publications

References 14 publications

Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience

Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience

3D CNN Based Automatic Diagnosis of ADHD Using fMRI Volumes

Deep Volumetric Ambient Occlusion

Contact Info

Product

Resources

About