Improvement of texture based image segmentation algorithm for HE stained tissue samples

Author

Windisch, Gergely ; Kozlovszky, Miklos

Author_Institution

John von Neumann Fac. of Inf., Obuda Univ., Budapest, Hungary

fYear

2013

fDate

19-21 Nov. 2013

Firstpage

273

Lastpage

279

Abstract

Superpixel algorithms are becoming a widely used method for many computer vision applications, and it could be used as a basis of image segmentation for digital microscopy images of HE stained tissue samples. Research results show that among the many superpixel methods SLIC yields the best results when it comes to boundary adherence accuracy for normal images. In an effort to find out if it can be used for segmenting tissue images we have devised a benchmark to measure the performance of SLIC and tried improving the performance by careful tuning of the parameters to better fit SLIC to our special image processing needs.

Keywords

biological tissues; computer vision; image segmentation; image texture; iterative methods; medical image processing; HE stained tissue samples; SLIC; computer vision; digital microscopy image; image processing; simple linear iterative clustering; superpixel algorithm; texture based image segmentation; Accuracy; Clustering algorithms; Gold; Image segmentation; Informatics; Shape; Standards; SLIC; Superpixels; tissue sample segmentation;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence and Informatics (CINTI), 2013 IEEE 14th International Symposium on

Conference_Location

Budapest

Print_ISBN

978-1-4799-0194-4

Type

conf

DOI

10.1109/CINTI.2013.6705205

Filename

6705205