2012 Fiscal Year Final Research Report

Cortical dysplasia detection from neonatal brain MR images usingintelligent medical image processing

Research Project

Project/Area Number	22680020
Research Category	Grant-in-Aid for Young Scientists (A)
Allocation Type	Single-year Grants
Research Field	Sensitivity informatics/Soft computing
Research Institution	University of Hyogo
Principal Investigator	KOBASHI Syoji 兵庫県立大学, 大学院・工学研究科, 准教授 (00332966)
Project Period (FY)	2010 – 2012
Keywords	難治てんかん / 皮質形成異常 / 核磁気共鳴画像 / ソフトコンピューティング / 計算機画像診断支援 / 知的画像処理
Research Abstract	Although subdural electrode has been used to detect cortical dysplasia, it cannot be applied to allpatients because of invasively . And, it cannot determine the boundary of cortical dysplasia legion. Thisstudy proposed a non-invasive cortical dysplasia detection and region segmentation method usingmagnetic resonance (MR) images. Firstly , this study proposed parenchyma segmentation method inbrain MR images based on fuzzy object model. Secondly , this study proposed a method of estimatingdegree cortical dysplasia using texture features and fractal dimension. The experimental results showedthe proposed method segmented the parenchymal region with a false-positive-volume-fraction of 0.75 %and detected cortical dysplasia with a mean sensitivity of about 90 %.