New forms of computational tomographic and super-resolution imaging
Conventional optical microscopy focuses on designing optical systems to faithly replicate an image of a magnified object to reveal very small spatial features. Image quality, and thus the ability to observe small spatial features, is limited by the ability to form high quality images for widefield microscopy, or for focusing to the smallest possible spot for laser scanning microscopy. Computational imaging can sidestep such limitations by taking into account a model of the image process, and this way computational imaging is able to produce high resolution imaging with a greater flexibility for optical systems. I will provide an overview of several recent results on computational imaging from my group: 1) three dimensional widefield second harmonic generation tomographic imaging based in defocused illumination, 2) tomographic fluorescent imaging by mimicking coherent light propagation, and 3) saturated super deconvolution microscopy.