poca-SMLM.github.io

An optimized implementation of DBSCAN is available in the Localizations//DBSCAN tabs.

DBSCAN uses two main parameters to cluster: a distance used to compute the localizations’ density (designed as Eps-neighborhood in the original paper) and a minimum number of points in this distance to classify localizations as core points (MinPts in the original paper).

When you click on the DBScan button, PoCA will compute the clustering with the current parameters. PoCA then compute the size of the different objects and display the results in a table and in an histogram. Beware: objects are not internally stored by PoCA. You need to create them explicit after through the appropriate button to have access to them in the ObjectList tab.

 Create objects with the current triangles/tetrahedra selected.

 Toggle rendering of the DBSCAN result.

Back to main page