Automatic 3D segmentation of human brain images using data-mining techniques

Author

Uher, Vaclav ; Burget, Radim

Author_Institution

Dept. of Telecommun., Brno Univ. of Technol., Brno, Czech Republic

fYear

2012

fDate

3-4 July 2012

Firstpage

578

Lastpage

580

Abstract

This paper proposes a method for automatic 3D segmentation of human brain CT scans using data mining techniques. The brain scans are processed in 2D and 3D. The proposed method has several steps - image pre-processing, segmentation, feature extraction from segments, data mining, and post-processing. The method introduced is implemented in 3D image processing extension for the RapidMiner platform, and both are provided as open source. With testing data the resultant performance selection of tissue slices from the brain image was 98.08% when compared to human expert results.

Keywords

biological tissues; computerised tomography; data mining; feature extraction; image segmentation; medical image processing; 3D image processing; RapidMiner platform; automatic 3D segmentation; computer tomography; data-mining techniques; feature extraction; human brain CT scans; human brain images; image postprocessing; image preprocessing; image segmentation; tissue slices; Brain; Feature extraction; Filtering algorithms; Image segmentation; Maximum likelihood detection; Nonlinear filters; Testing; 3D segmentation; CT; Image processing; RapidMiner; brain scan; data mining; open source; visualization;

fLanguage

English

Publisher

ieee

Conference_Titel

Telecommunications and Signal Processing (TSP), 2012 35th International Conference on

Conference_Location

Prague

Print_ISBN

978-1-4673-1117-5

Type

conf

DOI

10.1109/TSP.2012.6256362

Filename

6256362