Image analysis of arabidopsis trichome patterning in 4D confocal datasets

  • Authors:

  • R. Bensch;O. Ronneberger;B. Greese;C. Fleck;K. Wester;M. Hülskamp;H. Burkhardt

  • Affiliations:

  • Institute of Computer Science, University of Freiburg, Freiburg, Germany and Centre for Biological Signalling Studies, Freiburg, Germany;Institute of Computer Science, University of Freiburg, Freiburg, Germany and Centre for Biological Signalling Studies, Freiburg, Germany;Center for Biological Systems Analysis and Faculty of Biology, University of Freiburg;Center for Biological Systems Analysis, University of Freiburg;Botanical Institute III, University of Köln;Botanical Institute III, University of Köln;Institute of Computer Science, University of Freiburg, Freiburg, Germany and Centre for Biological Signalling Studies, Freiburg, Germany

  • Venue:
  • ISBI'09 Proceedings of the Sixth IEEE international conference on Symposium on Biomedical Imaging: From Nano to Macro
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this article, we present an approach for the automated extraction of quantitative information about trichome patterning on leaves of Arabidopsis thaliana. Time series of growing rosette leaves (4D confocal datasets, 3D + time) are used for this work. At first, significant anatomical structures, i.e. leaf surface and midplane are extracted robustly. Using the extracted anatomical structures, a biological reference coordinate system is registered to the leaves. The performed registration allows to determine intra- as well as inter-series spatio-temporal correspondences. Trichomes are localized by first detecting candidates using Hough transform. Then, local 3D invariants are extracted and the candidates are validated using a Support Vector Machine (SVM).