Research profile dr. M.H.F. (Michael )Wilkinson

I work on image analysis, signal processing, and computer vision techniques based on mathematical morphology, and in particular methods based on connectivity. Applications range from tracking subatomic particles in particle accelerators, through biomedical imaging and remote sensing,  to processing of large astronomical image and volume data sets. Developing new methods to model perceptual grouping of pixels into semantically important regions in ways that transcend regular connectivity and development of fast parallel algorithms are key parts of the research. Another important theme is the development of memory and computationally efficient algorithms for multi-scale analysis, along with integration of machine learning into early filtering stages of the image analysis pipeline.

Current projects include detection of merging galaxies in large surveys, in collaboration with the Kapteyn Astronomical Institute in Groningen, tracking subatomic particles using path-based operators in collaboration with KVI-CART in Groningen, development of high performance image processing algorithms for remote sensing in collaboration with the EC Joint Research Centre in Ispra, Italy, and development of computer vision techniques for a hedge-trimming robot in the Horizon-2020 Trimbot-2020 consortium.

Key words: mathematical morphology, connected filters, lattice theory, multi-scale image and signal analysis, machine learning, adaptive and trainable filters, parallel algorithms.

Websites:
Group website: http://www.cs.rug.nl/is/Main/HomePage
Employee (corporate) webpage: http://www.rug.nl/staff/m.h.f.wilkinson/
Personal homepage: http://www.cs.rug.nl/~michael