Affordable Access

Download Read

Classifying galaxies with mathematical morphology



In this report we present an investigation on whether methods and filters used in the field of mathematical morphology could be applied for classifying galaxies. Using multiscale connectivity and k-flat zone filtering we computed the differences in the pattern spectra of elliptical and spiral galaxies and of merging and projected galaxies. Based on the pattern spectra decision trees were built and tested using the C4.5 algorithm. With cross-validation but without bagging C4.5 classified around 60 to 75 % correct. The results show that classifying galaxies with mathematical morphology is an interesting subject for further research.

There are no comments yet on this publication. Be the first to share your thoughts.


Seen <100 times
Downloaded <100 times