Reconstruction of high-resolution atomic force microscopy measurements from fast-scan data using a Noise2Noise algorithm

• DOI: 10.1016/j.measurement.2024.114263
• Publication Date: 2024-02-02T05:57:16Z
• Authors: Eva Natinsky; Ryan M. Khan; Michael Cullinan; Rémi Dingreville

The acquisition of large atomic-force-microscopy (AFM) scans at nanoscale resolutions can take hours and produce datasets with millions pixels, which is time consuming computationally expensive to analyze. In this paper, we present an approach speed up process by using a computer-vision algorithm, namely the Noise2Noise reconstruct high-resolution, low scan AFM data from high-speed, noisy, sparsely sampled data. This algorithm trained on various noise types reproduce different sources experimental noises encountered during Our results demonstrate that sparse, uniform 20 × μm 128 pixel resolution be processed within seconds, output image comparable higher quality raw required 30 min or more collect, reducing not only analysis time, but also size being collected.