Open AccessAutomated Midline Setting for Brain Image Analysis to Detect Intracerebral Hemorrhage
Author(s) -
Fahmi Fahmi,
Sawaluddin Nasution,
. Sawaluddin
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1566/1/012011
Subject(s) - artificial intelligence , intracerebral hemorrhage , set (abstract data type) , lateralization of brain function , computer science , pixel , process (computing) , right hemisphere , computer vision , reading (process) , cerebral hemisphere , skull , psychology , medicine , radiology , audiology , surgery , subarachnoid hemorrhage , political science , law , programming language , operating system
One crucial step in detecting intra-cerebral hemorrhage automatically is to develop a midline setting that will divide the brain structure into 2 parts: the right hemisphere and the left hemisphere. This is useful for the process of detecting bleeding in the brain. By comparing the right hemisphere and the left hemisphere, bleeding detection can be done quickly and automatically on the application system. In this study, we proposed a method that was developed based on the pixel scan algorithm and the determination of the center of weight of the segmented part of the brain. The midline was set as the longest line possible drawn from edges of the brain skull. The result was then compared to manual reading by conducting user agreement surveys to radiology experts. The result shows that more than 75% of the cases experts strongly agree and agree with the midline setting automatically set by the algorithm, but 65% of them were dissatisfied and suggested improvement of the algorithm.