The phase-shift method in electron-off-axis holography: using neural network techniques

Abstract The common technique to extract amplitude and phase of the complex electron image wave from off-axis electron holograms is based on a band-pass filter that is applied to the spectrum of the hologram; therefore the carrier frequency of the hologram fringes has to be sufficiently high. The phase-shift method, which is well known from light-optical interferometry, uses three holograms each with a specific shift in the lateral position of the fringes to reconstruct amplitude and phase in real space, thus avoiding the need for fine hologram fringes. Neural networks are used for the reconstruction of the image wave and compared in performance with the analytical method.