Multiphoton microscopy at the intersection of super resolution imaging and artificial intelligence

Multiphoton excitation optical microscopy intrinsically allows multimodal fluorescence imaging allowing 3D, super resolved and life-time imaging including image scanning microscopy. Moreover intrinsic fluorescence and second harmonic generation mechanism of contrast produce labe-free imaging. Recently, Mueller matrix microscopy has been used to expand label-free capabilities. This permits to get multiple label-free images with simultaneous acquisition of the fluorescence counterpart. The availability of multimodal images from the very same sample is used to implement a paired image-to-image translation to train a generative adversarial network to produce a fluorescence image from its label-free twin. An interesting case study is related to understanding the visualization of chromatin organization.