Improving accuracy and uncertainty quantification of deep learning based quantitative MRI using Monte Carlo dropout
Image and Video Processing
2023-11-07 v2 Artificial Intelligence
Computer Vision and Pattern Recognition
Medical Physics
Abstract
Dropout is conventionally used during the training phase as regularization method and for quantifying uncertainty in deep learning. We propose to use dropout during training as well as inference steps, and average multiple predictions to improve the accuracy, while reducing and quantifying the uncertainty. The results are evaluated for fractional anisotropy (FA) and mean diffusivity (MD) maps which are obtained from only 3 direction scans. With our method, accuracy can be improved significantly compared to network outputs without dropout, especially when the training dataset is small. Moreover, confidence maps are generated which may aid in diagnosis of unseen pathology or artifacts.
Cite
@article{arxiv.2112.01587,
title = {Improving accuracy and uncertainty quantification of deep learning based quantitative MRI using Monte Carlo dropout},
author = {Mehmet Yigit Avci and Ziyu Li and Qiuyun Fan and Susie Huang and Berkin Bilgic and Qiyuan Tian},
journal= {arXiv preprint arXiv:2112.01587},
year = {2023}
}