Size Measurement of Nanoparticle Assembly Using Multilevel Segmented TEM Images

الملخص EN

Multilevel image segmentation is demonstrated as a rapid and accurate method of quantitative analysis for nanoparticle assembly in TEM images.

The procedure incorporating K-means clustering algorithm and watershed transform is tested on transmission electron microscope (TEM) images of FePt-based nanoparticles whose diameters are less than 5 nm.

By solving the nanoparticle segmentation and separation problems, this unsupervised method is useful not only in the nonoverlapping case but also for agglomerated nanoparticles.

Furthermore, the method exhibits scale invariance based on comparable results from images of different magnifications.