Fig. 1: Overview of the denoising method.
From: Utilizing the predictable binding kinetics of DNA-PAINT to denoise super-resolution images
Localizations are first clustered based on density by the DBSCAN algorithm followed by a k-means clustering step that partitions these clusters into localization-precision-sized clusters. For each cluster, a time trace representing the binding activity is constructed and the time between binding events is measured. A hypothesis test for exponentiality is performed on the distribution of the times between binding events and clusters with traces that pass the test are filtered as specific.
