Meeting Banner
Abstract #1334

Evaluation of input data and UNet based convolutional network architectures for automated muscle annotation in 2D and 3D

Martijn Froeling 1, Lara Schlaffke2, Marlena Rohm2, Ivana Isgum3, Hermien E Kan4, and Jelmer M Wolterink3

1Department of Radiology, University medical center utrecht, Utrecht, Netherlands, 2Department of Neurology BG, University Hospital Bergmannsheil, Ruhr-University Bochum, Bochum, Germany, 3Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands, 4Dept of Radiology, C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands

Manual annotation of muscle is still one of the most time-consuming steps in skeletal muscle MRI research. In this study we have investigated three aspects of automated muscle annotation using deep convolutional networks. First, we directly compare five different network architectures. Second, we compare the effect of providing various input data all based on Dixon imaging. Third, we investigate the effect of the amount of training data provided to the network. In summary we found that UNet-like convolutional networks allow for accurate and precise annotation of calf muscle in 2D and 3D and that the data provided is the strongest predictor of success.

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

Join Here