Optical microscopy is essential in biomedical and clinical research facilitating “seeing is believing” paradigm, capacitating diagnostically crucial examination of cells/tissues. It is under constant development with a goal of enhanced resolution, contrast, depth, speed, and information content, with computational techniques emerging recently as capable solutions (e.g., deep learning). Fluorescence microscopy is the most impactful microscopic technique. It relies on labeling the structure of interest with fluorophores – nanoscopic light-bulbs attached to the specimen rendering its high-contrast images. Despite their gigantic success, all fluorescence methods have inherent limitations: (1) time-consuming fluorescence labeling is prone to phototoxicity and (2) it alters normal physiological processes invasively affecting natural cell-behavior jeopardizing the reliability of the study. Thus, a strong demand on capable label-free methods arises with a goal of studying the dynamics of live cells and their natural physiological activities. The sample is not chemically stained and therefore can be imaged basing on its endogenous contrast agents, i.e., internal absorption, scattering or refractive index (phase delay). Zernike/Nomarski phase contrast microscopy are well-known marker-free methods, but they suffer from halo and low contrast and do not provide quantitative phase information, thus are not suitable for standardized analysis. Quantitative Phase Microscopy (QPM) stands out among modern label-free imaging techniques as extremely capable high-contrast approach based on the interference measurement of sample refractive index distribution. Although application-oriented research provides exciting advancements, QPM benchtop devices tend to be bulky and costly, both in financial terms and in effort required to master them. They are based on a modified brightfield microscope, thus possess limitations in terms of finite microscope objective numerical aperture and small depth of focus.  

These disadvantages are bypassed by the lensfree holographic microscopy (LHM), which in easy-to-assemble manner (light source+specimen+camera) captures a “holographic shadow” of the illuminated object. This “shadow” is a Gabor in-line hologram generated upon interference of non-scattered (reference) and object scattered wave-fronts. Numerical reconstruction opens up possibilities to study extremely large volumes (thousands of cubic millimeters) without the depth of focus limitations, which is a fantastic advantage of LHM. The problem to be solved within the GaboScope is connected with the fact that classical schemes employed to recover, upon backpropagation, the in-focus objects are very sensitive to low-quality of holograms (strong noise and incoherent background, low contrast, high sample density and scattering etc.) and apply time-consuming algorithmic/experimental solutions which impede deterministic high-throughput marker-free investigation of dynamic processes in a large holographically rendered volume.