Generative resolution-enhanced microscopy based on computational stitching of Fourier spectra

Scanning Microscopies such as Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) are widely used in nanotechnology to display (Kalinin et al., 2023) and quantify surface nanoroughness. However, their measurements suffer from scale and spatial limitations which hinder their full exploitation in the optimization of nanostructure fabrication and performance. The aim of this work is to overcome these limitations by proposing a computational method which enables the generation of multiple surface measurements (overcome spatial limitation) with enhanced scale content and increased resolution (overcome scale limitation) successfully mimicking the real surfaces. The method exploits the benefits of Fourier transform and is easily implemented based on Fourier spectra stitching. It can provide astonishing increase of resolution (>10 times) in a large variety of image textures and, is fast with no requirements for training with large data sets. The so-called generative Fourier Spectra Stitching (gFSS) method is validated in synthesized rough surfaces and applied in real experimental AFM and SEM images of magnetic and metal surfaces with very promising results.