Free Access
Issue
Microsc. Microanal. Microstruct.
Volume 5, Number 2, April 1994
Page(s) 105 - 119
DOI https://doi.org/10.1051/mmm:0199400502010500
Microsc. Microanal. Microstruct. 5, 105-119 (1994)
DOI: 10.1051/mmm:0199400502010500

Automated segmentation algorithm for electron microscope images with strong grey level fluctuations

Xin Zheng et Max Kolb

Laboratoire de Chimie Théorique, Ecole Normale Supérieure de Lyon, 46 allée d'Italie, 69364 Lyon cedex 07, France


Abstract
An automated algorithm for image segmentation based on nucleation and growth of the background is presented. The graytone parameters of the algorithm are chosen by the algorithm itself. Only a single graytone cutoff remains to be chosen by hand. This cutoff is more general than the graytone parameters themselves, as it can be fixed for a whole class of images. Different ways to implement the automated choice of the graytone parameters are investigated. The role of the geometrical parameters of the automated algorithm is also discussed.

PACS
0778 - Electron, positron, and ion microscopes; electron diffractometers.

Key words
Algorithms -- Segmentation -- Images -- Electron microscopy -- Grey level image -- Fluctuations -- Particles -- Metals -- Sintering -- Metrology -- Physics -- Artificial intelligence -- Applied sciences -- Information theory -- Signal processing -- Connectionism


© EDP Sciences 1994