Topological Watershed

From LRDE

The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

Abstract

Dividing a picture into area of interest is called picture segmentation, it is useful in particular to point out cancerous cells in medical imaging. The emphWatershed Transform provides such a segementation and can be implemented in many ways. Here we will focus on the emph Topological Watershed, a performant algorithm producing results with nice properties. In this report, we will show how this algorithm had been implemented in Milena, the C++ generic image processing library of Olena, developed at the LRDE. We will first see how to treat usual image format and then generalize it to trickier formats like pictures mapped on general graphs.