Anna introduced a new mathematical approach for dimensionality reduction that we incorporated into loss functions to augment target information and improve performance.
We used epistemic uncertainty estimates to discover potential abnormalities in diseased OCT scans. The uncertainty maps are obtained by a Bayesian U-Net trained on healthy OCT scans with weak labels of the retinal layers.
We present an extensive description and evaluation of our method for blood vessel segmentation in fundus images based on a discriminatively trained fully connected conditional random field model.