Seminar “Task of localization primary signs of a brain stroke”
15 September 15 at 16:30 MSK
This report will tell about the way to solve the problem of localization of the primary signs of a brain stroke. The stages of data preprocessing, filtering, clustering are touched upon, methods for assessing damage are proposed. The task of finding the primary signs of stroke is important due to the subjectivity of visual analysis by a neurologist.
As a result, when making decisions related to medical manipulations, an independent algorithmic assessment of the patient’s condition is needed. Usually, machine learning methods are used to solve such problems, but the specifics of the disease and its manifestations, as well as complex access to data, force us to think through approaches that bypass the use of machine learning in favor of more conservative methods.
Aleksey Lobanov, Phd student PFUR, engineer of Laboratory 81 of Institute of Control Sciences.