Abstract: A microscope system (100) configured to record images in at least a first and a second imaging mode (501, 502), comprising: An objective (1) collecting light (201) from a sample (11), An illumination module coupled to the objective, A first reimaging objective (5) generating an intermediate image of the sample and a second reimaging objective (6) that relays the intermediate image onto a detection module, An evaluation module (200) comprising a machine learning method (DL), trained with a first and a second set of images of the same sample, wherein the first and second set has been acquired in the first (501) and second imaging mode (502), respectively, wherein upon acquisition of an image (400) in the second imaging mode (502) the trained machine learning method (DL) outputs a restored image (401) that comprises fewer aberrations than the image (400) acquired in the second imaging mode (52, 53, 57).

Abstract: The invention relates to a sample holder assembly (1) for microscopy comprising at least the following components: A first member (10) having a first opening (14) on a first side of the first member (10), and a circumferential wall portion (12) having an inward facing side (12-1) that faces toward the first volume (V1) and an outward facing side (12-2) opposite the inward facing side (12-1); A second member (20) having a first opening (24) at a first side of the second member (20), and a circumferential wall portion (22) having an inward facing side (22-1) facing toward the second volume (V2), wherein a shape of the outward facing side (12-2) of the wall portion (12) of the first member (10) is complementary to the inward facing side (22-1) of the wall portion (22) of the second member (20), such that the wall portion (12) of the first member (10) can be inserted in the second volume (V2), such that the first volume (V1) is at least partially comprised by the second volume (V2); A transparent membrane (30) ha