Hyperight

Deep Learning Without A Big Dataset: Segmenting Brain Tumors – Lars Sjösund

Deep learning achieves very good results in the task of segmenting brain tumors, even when the available training dataset is quite small. We look at why this is, and how we can create an algorithm that is on par with human experts.

Key Points:

-Do we always need a large dataset when working with deep learning?

– How to design a deep learning algorithm for segmentation?

– How to handle imbalanced data?

Add comment