Modelling of Learning and Perception

Centre of Excellence in Computational Complex Systems Research


Computer Vision for Electron Tomography

Researchers: Sami Brandt, Vibhor Kumar, Jukka Heikkonen, and Peter Engelhardt

In structural biology, electron tomography is used in reconstructing three-dimensional objects such as macromolecules, viruses, and cellular organelles to learn their three-dimensional structures and properties. The reconstruction is made from a set of transmission electron microscope (TEM) images which may be obtained by tilting the specimen stage by small angular increments (single axis tilting). In order to successfully perform the 3D reconstruction in electron tomography, transmission electron microscope images have to be accurately aligned or registered. The alignment problem can be posed as a motion estimation problem that can solved by using geometric computer vision methods.

Previously, we have developed two methods where the registration is automated. Most accurate alignment can be achieved if conventional colloidal gold markers are used. In contrast to the manual picking, our method collects the gold beads automatically by using recent techniques of computer vision. For cases when it is not possible to use gold particles, we have proposed an alternative method that is based on tracking high curvature points of the intensity surface of the images. Results show almost as good performance as we have obtained by using fiducial markers (Figure 1). The development of the alignment algorithms is still going on for better accuracy and to take computational aspects into consideration.

Figure 1

Figure 1. Stereo image pair of a reconstructed microvillus where the image series has been aligned by tracking certain interest points of the image intensity surface.