Bayesian Generative Models for Knowledge Transfer in MRI Semantic Segmentation Problems


During the MIDL2020, I and Anna Kuzina presented our work on transfer learning for 3D MRI segmentation. How to learn deep and big 3D UNet-like model with only 5 or 10 training images? It is a common question for medical data. We address it by imposing prior distribution over the convolutional kernels as the generative model (VAE)!

Thanks to the MIDL, here it is the page with video, slides and short paper. If you are interested, there is published full paper. If you decide to try this approach or develop it, we published the code.

1-minute presentation:

Abstract of the paper:

Automatic segmentation methods based on deep learning have recently demonstrated state-of-the-art performance, outperforming the ordinary methods. Nevertheless, these methods are inapplicable for small datasets, which are very common in medical problems. To this end, we propose a knowledge transfer method between diseases via the Generative Bayesian Prior network. Our approach is compared to a pre-train approach and random initialization and obtains the best results in terms of Dice Similarity Coefficient metric for the small subsets of the Brain Tumor Segmentation 2018 database (BRATS2018).