Published on Mon Mar 21 2016

Variational Autoencoders for Feature Detection of Magnetic Resonance Imaging Data

R. Devon Hjelm, Sergey M. Plis, Vince C. Calhoun

Independent component analysis (ICA) has become the de-facto standard in many medical imaging settings. The limitation of ICA to square linear transformations has not been overcome. As an alternative, we present feature analysis in medical imaging as a problem solved by Helmholtz machines.

0
0
0
Abstract

Independent component analysis (ICA), as an approach to the blind source-separation (BSS) problem, has become the de-facto standard in many medical imaging settings. Despite successes and a large ongoing research effort, the limitation of ICA to square linear transformations have not been overcome, so that general INFOMAX is still far from being realized. As an alternative, we present feature analysis in medical imaging as a problem solved by Helmholtz machines, which include dimensionality reduction and reconstruction of the raw data under the same objective, and which recently have overcome major difficulties in inference and learning with deep and nonlinear configurations. We demonstrate one approach to training Helmholtz machines, variational auto-encoders (VAE), as a viable approach toward feature extraction with magnetic resonance imaging (MRI) data.

Thu Jan 25 2018
Machine Learning
Development of ICA and IVA Algorithms with Application to Medical Image Analysis
Independent component analysis (ICA) is a widely used BSS method. It can only achieve source recovery, subject to scaling and permutation, through the assumption of statistical independence. Independent vector analysis (IVA) extends the applicability of ICA by jointly decomposing multiple datasets.
0
0
0
Wed Feb 26 2020
Machine Learning
ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICA
In our model family, the energy function is the product between two feature extractors. We show that under mild conditions, the features are unique up to scaling and permutation.
0
0
0
Wed Nov 27 2019
Machine Learning
High- and Low-level image component decomposition using VAEs for improved reconstruction and anomaly detection
Variational Auto-Encoders have often been used for unsupervised pretraining, out-of-distribution and anomaly detection. VAEs often lack the ability to produce sharp images and learn high-level features. We propose to alleviate these issues by adding a new branch to conditional
0
0
0
Sun Dec 16 2018
Machine Learning
Fast MVAE: Joint separation and classification of mixed sources based on multichannel variational autoencoder with auxiliary classifier
This paper proposes an alternative algorithm for multichannel variational autoencoder (MVAE) The proposed method employs an auxiliary classifier VAE, an information-theoretic extension of the conditional VAE.
0
0
0
Thu Jun 25 2020
Machine Learning
Neural Decomposition: Functional ANOVA with Variational Autoencoders
Variational Autoencoders (VAEs) have become a popular approach for dimensionality reduction. Despite their ability to identify latent structures embedded within high-dimensional data, these latent representations are typically hard to interpret on their own.
0
0
0
Mon May 03 2021
Machine Learning
Fusing multimodal neuroimaging data with a variational autoencoder
0
0
0