Comparison of Data-Driven and Morphological Features for Cell Segmentation in Histopathological Images


KARAASLAN Ö. F. , BİLGİN G.

IEEE 29th Signal Processing and Communications Applications Conference (SIU'21), İstanbul, Turkey, 9 - 11 June 2021, vol.1, pp.1-4 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1
  • Doi Number: 10.1109/siu53274.2021.9477704
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.1-4
  • Keywords: Histopathological image analysis, adaptive data analysis, empirical mode decomposition, variational mode decomposition, extended morphological profiles

Abstract

In this study, it is aimed to increase the segmentation performance of the cells in digital histopathological images by morphological feature extraction methods. For this purpose, it is proposed to evaluate the features extracted by the extended morphological profiles method. First, principal component analysis of the images in the RGB color space of digital histopathological images is performed, then the features are extracted using the extended morphological profiles method. Then, a feature set is created from these attributes and classified with support vector machines, which are kernel based classifiers. The results are compared and evaluated according to three different metrics with other results that were previously obtained in the same data set. In the application results section, the results obtained in this study are presented in full detail.