Artificial Enhanced Image Inpainting for Broadband SMM

Scanning probe microscopy techniques changed the understanding of micro and nanometer scaled objects. Since the invention of the scanning tunneling microscope and the atomic force microscope, these techniques were further developed and combined with different measurement methodologies such as scanning microwave microscopy. By introducing an electromagnetic signal in the microwave frequency range via the scanning probe microscope tip to the sample, electromagnetic characteristics of the specimen can be acquired. Using a series of frequencies in one scan, even more information like dopant concentration or different layers of material can be captured. The drawback of this method is the long data acquisition time, which cannot be neglected on nanometer scale. To address this issue, we are using reduced raster scan techniques with broadband scanning microwave microscopy and calculate the missing data information by using image inpainting algorithms afterwards. In this work we will present the development of a deep learning algorithm and compare the output of it with different inpainting algorithms, which can even be used in other scanning probe microscope setups.