Chapter 7
Alignment of Tilt Series
A. Verguet, C. Messaoudi, C. O. S. Sorzano and S. Marco
Abstract Computing of three-dimensional reconstructions from images obtained
by transmission electron tomography needs three main steps: data acquisition,
projection alignment, and 3D reconstruction. In this chapter we will focus on the
process of alignment moving from the justification of its need to the study of the
different classical approaches (cross-correlation, use of added fiducial markers) that
have been commonly used in this alignment process. We will also discuss the most
recent algorithms (multiscale registration, invariant feature recognition) as they
have been adapted to Electron Tomography and improved to increase the accuracy
and resolution of the final tomograms.
7.1 Why Do We Need to Align Tilt Series?
We may consider that tomography consists in determining a numerical representation in three dimensions of an object from their projections which requires the
mathematical combination of the acquired projections by a process called reconstruction. In the case of transmission electron tomography (TET), these projections
A. Verguet
INSERM U1196, Centre Universitaire, Bat. 112, 91405 Orsay Cedex, France
e-mail: amandine.verguet@ssji.net
C. Messaoudi
CNRS UMR9187, Centre Universitaire, Bat. 112, 91405 Orsay Cedex, France
e-mail: cedric.messaoudi@curie.fr
C. O. S. Sorzano
Biocomputing Unit, National Center of Biotechnology (CSIC). C/Darwin,
3, Campus Universidad Autonoma (Cantoblanco), 28049 Madrid, Spain
e-mail: coss@cnb.csic.es
S. Marco (&)
Institut Curie, Centre de Recherche, Centre Universitaire,
Bat. 112, 91405 Orsay Cedex, France
e-mail: sergio.marco@inserm.fr
© Springer International Publishing AG 2018
E. Hanssen (ed.), Cellular Imaging, Biological and Medical Physics,
Biomedical Engineering, https://doi.org/10.1007/978-3-319-68997-5_7
183
Alignment of Tilt Series
A. Verguet, C. Messaoudi, C. O. S. Sorzano and S. Marco
Abstract Computing of three-dimensional reconstructions from images obtained
by transmission electron tomography needs three main steps: data acquisition,
projection alignment, and 3D reconstruction. In this chapter we will focus on the
process of alignment moving from the justification of its need to the study of the
different classical approaches (cross-correlation, use of added fiducial markers) that
have been commonly used in this alignment process. We will also discuss the most
recent algorithms (multiscale registration, invariant feature recognition) as they
have been adapted to Electron Tomography and improved to increase the accuracy
and resolution of the final tomograms.
7.1 Why Do We Need to Align Tilt Series?
We may consider that tomography consists in determining a numerical representation in three dimensions of an object from their projections which requires the
mathematical combination of the acquired projections by a process called reconstruction. In the case of transmission electron tomography (TET), these projections
A. Verguet
INSERM U1196, Centre Universitaire, Bat. 112, 91405 Orsay Cedex, France
e-mail: amandine.verguet@ssji.net
C. Messaoudi
CNRS UMR9187, Centre Universitaire, Bat. 112, 91405 Orsay Cedex, France
e-mail: cedric.messaoudi@curie.fr
C. O. S. Sorzano
Biocomputing Unit, National Center of Biotechnology (CSIC). C/Darwin,
3, Campus Universidad Autonoma (Cantoblanco), 28049 Madrid, Spain
e-mail: coss@cnb.csic.es
S. Marco (&)
Institut Curie, Centre de Recherche, Centre Universitaire,
Bat. 112, 91405 Orsay Cedex, France
e-mail: sergio.marco@inserm.fr
© Springer International Publishing AG 2018
E. Hanssen (ed.), Cellular Imaging, Biological and Medical Physics,
Biomedical Engineering, https://doi.org/10.1007/978-3-319-68997-5_7
183