Over the past 5 years there has been an increase in the use of convolutional neural networks in a broad variety of medical imaging applications. This is due in part to the increase in their popularity since their success in the 2012 ImageNet competition, but is also due to their adaptability across a range of medical imaging applications. These applications vary greatly; from the segmentation of knee cartilage to the detection of Alzheimer's disease in MRIs and much more. In this paper we will go over some of the cutting edge techniques being used specifically for the tasks of brain segmentation; classifying with both binary segmentation on brain lesions and hierarchical segmentation with tumors. The results are proving to be quite promising with many of the described techniques outscoring previous state-of-the-art systems.
"Convolutional Neural Networks in Medical Imaging,"
Scholarly Horizons: University of Minnesota, Morris Undergraduate Journal: Vol. 4
, Article 3.
Available at: http://digitalcommons.morris.umn.edu/horizons/vol4/iss2/3