We present a computational method, termed Wasserstein-induced ux (WIF), to robustly quantify the accuracyof individual localizations within a single-molecule localization microscopy (SMLM) dataset without groundtruthknowledge of the sample. WIF relies on the observation that accurate localizations are stable with respectto an arbitrary computational perturbation. Inspired by optimal transport theory, we measure the stabilityof individual localizations and develop an efficient optimization algorithm to compute WIF. We demonstratethe advantage of WIF in accurately quantifying imaging artifacts in high-density reconstruction of a tubulinnetwork. WIF represents an advance in quantifying systematic errors with unknown and complex distributions,which could improve a variety of downstream quantitative analyses that rely upon accurate and precise imaging.Furthermore, thanks to its formulation as layers of simple analytical operations, WIF can be used as a lossfunction for optimizing various computational imaging models and algorithms even without training data.
展开▼
