Sub-diffraction error mapping for localization microscopy images

Abstract

Assessing the quality of localization microscopy images is highly challenging due to difficulty in reliably detecting errors in experimental data, with artificial sharpening being a particularly common failure mode of the technique. Here we use Haar wavelet kernel analysis (HAWK), a localization microscopy data analysis method which is known to give results without artificial sharpening, to generate a reference image. This enables the mapping and quantification of this common artefact. By suppressing intensity information, we are able to map sharpening errors in a way which is not influenced by nonlinearity in the localisation imaging process. The HAWK Method for the Assessment of Nanoscopy (HAWKMAN) is a general approach which allows the reliability of localization information to be assessed.