Improving accuracy and uncertainty quantification of deep learning based quantitative MRI using Monte Carlo dropout
This work addresses the need for more reliable and interpretable deep learning models in medical imaging, particularly for quantitative MRI with limited data, though it is incremental as it builds on existing dropout techniques.
The paper tackled the problem of improving accuracy and uncertainty quantification in deep learning for quantitative MRI by using Monte Carlo dropout during both training and inference, averaging predictions to enhance results. The method significantly improved accuracy for fractional anisotropy and mean diffusivity maps from limited 3-direction scans, especially with small training datasets, and generated confidence maps for diagnostic aid.
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.