Open-source tools enable accessible and advanced image scanning microscopy data analysis

Alessandro Zunino, Eli Slenders, Francesco Fersini, Andrea Bucci, Mattia Donato, and Giuseppe Vicidomini (see publication in Journal )


Thanks to the advent of high-performance detector arrays, image scanning microscopy (ISM) has transformed the field of fluorescence laser-scanning microscopy1. In particular, the development of fast, sensitive single-photon avalanche diode (SPAD) arrays2 has paved the way for robust ISM imaging with exceptional information content3. In ISM, a focused excitation beam scans the sample. For each scan point, the SPAD array collects a small image — called a micro-image — of the diffraction-limited fluorescent region. In effect, each SPAD of the detector acts as an independent and displaced pinhole, generating a confocal-like image. Thus, the ISM dataset can be seen either as a large set of small wide-field images or as a small set of large confocal images, both point-of-views opening unique ways to process the data. Tailored algorithms enable the transformation of the raw ISM dataset into a single image with enhanced spatial resolution, signal-to-noise ratio, and optical sectioning. With the recent introduction into the market of commercial microscopes equipped with a detector array — namely, AiryScan from Zeiss, PRISM from Genoa Instruments, MATRIX from Abberior, and NSPARC from Nikon — we expect ISM to become increasingly widespread in use. With this in mind, we have designed an open-source tool called BrightEyes-ISM to provide the scientific community with a computational toolbox for processing the ISM datasets with cutting-edge algorithms for image reconstruction, image quality evaluation, and simulation. BrightEyes-ISM is an open-source Python package with three different modules, each containing functions specific to data loading, data analysis, or optical simulations. The simulation module contains a point spread function (PSF) simulator (Fig. 1a–c) and a tubulin simulator. The PSF simulator is currently built upon PyFocus4, a Python package that enables the vectorial calculation of focused optical fields. The simulator generates a set of PSFs, one for each element of the detector array, with all of the relevant physical parameters that can be fine-tuned. We also built a phase calculator that enables the introduction of any aberration or a user-defined phase mask. Optionally, the simulator can introduce a third field for stimulated emission depletion (STED), enabling the simulation of STED-ISM PSFs. The dataIO module enables the user to easily read and store data and metadata in a hierarchical data format (HDF), which is particularly convenient for ISM datasets that are inherently large and multi-dimensional (Fig. 1d). The analysis module contains libraries for state-of-the-art analysis of ISM data processing. The APR library contains the functions needed to perform adaptive pixel reassignment (APR), a fast and robust method to reconstruct a super-resolution image from an ISM dataset5 (Fig. 1e). The deconvolution library contains many functions to perform image deconvolution, the most important one being the multi-image deconvolution6 that enables an excellent alternative to APR for image reconstruction (Fig. 1f). The Focus-ISM library contains all of the functions needed to perform Focus-ISM, namely, an innovative algorithm that greatly enhances the optical sectioning of ISM, effectively removing the out-of-focus light from the ISM datasets7 (Fig. 1g). The FRC library provides the tools needed to perform Fourier ring correlation (FRC) analysis on the reconstruction, enabling a quantitative method to evaluate the resolution of the images8 (Fig. 1h). Notably, we also use FRC as a tool to estimate the PSFs from experimental data and perform blind multi-image deconvolution9. BrightEyes-ISM can be found on GitHub (, together with the complete documentation. Importantly, we intended our package to be useful both for developers and microscopy users. We therefore integrated the majority of the functionalities of BrightEyes-ISM into a Napari plugin, named Napari-ISM ( Napari is a multi-dimensional image viewer that provides a quick and easy way to explore and analyse the ISM data.