Morphological classification of dense objects in atom probe tomography data

The technique of atom probe tomography is often used to image solute clusters and solute atom segregation to dislocation lines in structural alloys. Quantitative analysis, however, remains a common challenge. To address this gap, we combined a cluster finding algorithm, a skeleton finder algorithm, and morphological classification of dense objects to distinguish solute clusters from solute-decorated dislocation lines, both being characterized by high solute atom densities. The proposed workflow is packaged into a graphical user interface available through GitHub. We illustrate its application on a synthetic dataset containing known objects and apply it to an experimental dataset obtained from a proton-irradiated Alloy 625 that contains high densities of Si-decorated dislocations and Si-rich clusters.