Meeting Banner
Abstract #3523

Transfer learning and convolutional neural net fusion for motion artefact detection

Christopher Kelly1, Max Pietsch1, Serena Counsell1, and J-Donald Tournier1

1Centre for the Developing Brain, King's College London, London, United Kingdom

The first step in a diffusion MRI pre-processing pipeline typically involves the manual removal of heavily motion-corrupted volumes. However, this process is both time consuming and potentially subjective. We propose to automate this process by training multiple deep convolutional neural networks (CNNs) and decision trees to achieve near human-level accuracy for rejection of outliers.

This abstract and the presentation materials are available to members only; a login is required.

Join Here