ABSTRACT

Alzheimer’s disease is an incurable progressive neurological brain disorder. Alzheimer’s disease disrupts the communication among neurons, resulting in loss of function and cell death [1]. In Alzheimer’s disease (AD), unusual proteins build up in and around neurons in the neocortex and hippocampus—parts of the brain that control memory. When these neurons die, people lose their capacity to remember and their ability to do everyday tasks [2]. The normal behavior of human beings depends on the functionality of the hippocampus. Manual segmentation by a specialist on the hippocampus takes many hours. In deep learning, there are various techniques available for segmentation process. In this chapter, a review is done on the different deep learning approaches to segmentation. First, we review the current deep learning approaches used for segmentation of anatomical brain structures and brain lesions with the help of white matter hyperintensities (WMH). Studies have found a direct association between total volume of WMH and increased odds for having AD. Finally, we provide a critical assessment of the current state and identify likely future developments and trends.