I just spent most of the past few years building my own motorized microscope for automated acquisition. As of yesterday, I successfully trained a tardigrade object detector and used it to do real-time tracking of tardigrades as they move around on a slide, for hours (yes, there is actually a hypothesis).
