"A framework for evaluating the performance of SMLM cluster analysis algorithms."
A critical contribution, #SMLM (#SingleMoleculeLocalizationMicroscopy) is highly confounded data, without validation, large scale (1e6+ 3D points), with #fuzzy localizations, non-specific labelling, multiple localizations, ... . Resolving protein complexes/ structures from such data is non-trivial to say the least. Robust evaluation of such methods (and p-choices) is welcome.