Variational Autoencoders for Feature Detection of Magnetic Resonance Imaging Data
This work addresses feature detection in medical imaging, specifically for MRI data, and appears incremental as it applies an existing method (VAE) to a new domain.
The paper tackles the problem of feature extraction in medical imaging by proposing variational autoencoders (VAEs) as an alternative to independent component analysis (ICA), demonstrating their viability for MRI data.
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.