Automation segmentation of PET image for brain tumors

Author

Zhu, Wanlin ; Jiang, Tianzi

Author_Institution

Nat. Lab. of Pattern Recognition, Chinese Acad. of Sci., Beijing, China

Volume

4

fYear

2003

fDate

19-25 Oct. 2003

Firstpage

2627

Abstract

The paper presents an improved fuzzy c-means (FCM) algorithm for obtaining segmentation results of PET image. The segmentation of images with low resolution is usually more difficult than images with high resolution on account of boundary definition difficulties. In order to extract tumor from a PET image, we have to specify the numbers of clusters and which may vary from one image to another when we apply FCM algorithm. However we can divide all contents of image into two parts: background and foreground. Then iterative fuzzy clustering was used and we can get desired results via parameters assessment. The advantage of the algorithm is completely automatic and simple. It is shown that the algorithm is robust for a lot of different datum by experiment.

Keywords

brain; image segmentation; iterative methods; medical computing; medical image processing; positron emission tomography; tumours; PET image; automation segmentation; background image content; boundary definition; brain tumors; cluster number; foreground image content; fuzzy c-means algorithm; high resolution; image segmentation; iterative fuzzy clustering; low resolution; parameter assessment; robust algorithm; Automation; Biomedical imaging; Clustering algorithms; Image resolution; Image segmentation; Iterative algorithms; Magnetic resonance imaging; Neoplasms; Optical imaging; Positron emission tomography;

fLanguage

English

Publisher

ieee

Conference_Titel

Nuclear Science Symposium Conference Record, 2003 IEEE

ISSN

1082-3654

Print_ISBN

0-7803-8257-9

Type

conf

DOI

10.1109/NSSMIC.2003.1352428

Filename

1352428