Volta

3DCAE-MRI: Overcoming Data Availability in Small Sample Size MRI Studies

<title>Abstract</title> Deep learning (DL) are data-driven models that learn abstract, hierarchical features from raw or low-processed data. These models perform exceptionally in image recognition tasks using large amounts of data. Thus, the availability of open-source large-scale annotated datasets is crucial to achieve high-performance models. Nevertheless, the usage of DL models in neuroimaging is restricted due to data availability. Transfer learning has been used to overcome the issue, nonetheless most studies focus on transferring knowledge from or to an Alzheimer's Disease context. In this paper, we propose the 3D Convolutional Autoencoder Magnetic Resonance Imaging (3DCAE-MRI), an unsupervised model, to facilitate the usage of supervised DL models in small sample size Magnetic Resonance Imaging (MRI) studies. We exploit different open-source MRI databases …

Reference & gravity metrics

Secondary-market trade history

References (0)

Cited by (1)