SEM3De: image restoration for FIB-SEM

Nom de la revue
Bioinformatics Advances
Rayane Hamdane Serir, Aurelie Deliot, Caroline Kizilyaprak, Jean Daraspe, Christine Walczak, Françoise Canini, Amandine Leleu, Sergio Marco, Frederic Ronzon, Cedric Messaoudi
Abstract

Abstract

Motivation
FIB-SEM (Focused Ion Beam—Scanning Electron Microscopy) is a technique to generate 3D images of samples up to several microns in depth. The principle is based on the alternate use of SEM to image the surface of the sample (a few nanometers thickness) and of FIB to mill the surface of the sample a few nanometers at the time. In this way, huge stacks of images can thus be acquired.
Although this technique has proven useful in imaging biological systems, the presence of some visual artifacts (stripes due to sample milling, detector saturation, charge effects, focus or sample drift, etc.) still raises some challenges for image interpretation and analyses.

Results
With the aim of meeting these challenges, we developed a freeware (SEM3De) that either corrects artifacts with state-of-the-art approaches or, when artifacts are impossible to correct, enables the replacement of artifactual slices by an in-painted image created from adjacent non-artifactual slices. Thus, SEM3De improves the overall usability of FIB-SEM acquisitions.

Availability and implementation
SEM3De can be downloaded from https://sourceforge.net/projects/sem3de/ as a plugin for ImageJ.