Cross-correlation based detection of nanoparticles in SEM images from sedimentation cell experiments

M. Simon, E. M. Wülfers, Antoine Tavernier, S. Fritsch-Decker, E. Müller, J. Seiter, C. Weiss, O. Dössel, D. Gerthsen, G. Seemann

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

To meet the increasing demand for reliable and fast experimental methods for nanoparticle investigation, we developed a combination of a scanning electron microscopy (SEM) imaging procedure and a computer program for counting nanoparticles (SiO2) on cells (A549) in sedimentation experiments. Our method is based on cross-correlation and is called "template-based cross-correlation". We have tested this method of nanoparticle detection regarding its ability to quantify nanoparticles on cells in SEM images of sedimentation experiments. Our iterative template-based algorithm starts with a given disk template and generates a template of the nanoparticles via cross-correlation. After multiple additional cross-correlations for improving the template, the algorithm performs a final cross-correlation with the most improved template. We found that this special kind of algorithm is useful to detect single particles on plain surfaces but fails on agglomerated particles or complicated background scenarios, e.g. cell-surfaces. Due to this, advanced programs for analyzing and quantifying nanoparticles in SEM images have to use multiple methods to bridge the gaps between the particular shortcomings of the single methods. Furthermore, a semi-automatic program design with the ability of interactive correction is probably the most reliable way to deal with the challenging demands of nanoparticle analysis in SEM images.

Original languageEnglish
Pages (from-to)S510-S513
JournalBiomedizinische Technik
Volume59
DOIs
StatePublished - 1 Oct 2014
Externally publishedYes

Fingerprint

Dive into the research topics of 'Cross-correlation based detection of nanoparticles in SEM images from sedimentation cell experiments'. Together they form a unique fingerprint.

Cite this