News |  Sitemap |  Contact
PDF Export  | 

Fully automatic landmark extraction for ImageJ


Landmark correspondences can be used for various tasks in image processing. Examples of use include automatic image alignment, reconstruction of panoramic photographies, object recognition and simultaneous localization and mapping for mobile robots. The computer vision community knows several techniques for extracting and pairwise associating such landmarks using distinctive invariant local image features. Two very successful methods are the Scale Invariant Feature Transform (SIFT) by David Lowe (2004) and the Multi-Scale Oriented Patches (MOPS) by Matthew Brown (2005).

We implemented these methods in the Java programming language for seamless use in ImageJ. We use it for fully automatic registration of gigantic serial section Transmission Electron Microscopy mosaics. Using automatically extracted landmark correspondences, the registration of large image mosaics simplifies to globally minimizing the displacement of corresponding points.

In this talk, we will give an introduction to automatic landmark extraction and demonstrate our implementation for ImageJ. Furthermore we will outline our registration procedure as an example of what it can be used for.


sift, scale invariant feature transform, automatic landmark detection, registration, mosaic, transmission electron microscopy

Stephan Saalfeld

Max Planck Institute of Molecular Cell Biology and Genetics Dresden


Short Biography
Stephan Saalfeld is a PhD student in the Max Planck Institute of Molecular Cell Biology and Genetics Dresden. He studied Media Computer Science in the Technische Universitaet Dresden and got his Diploma in 2008. He is focused on Image processing in Confocal and Electron Microscopy.

© Luxembourg Institute of Science and Technology | Legal Notice