DANIEL FREEDMAN

We present two fundamental contributions to tackle the problem of high-density overlapping PSFs for accurate 3D localization. First, we employ a convolutional neural network (CNN) for 3D localization from dense fields of overlapping emitters with the Tetrapod PSF. Second, we engineer the optimal PSF for high-density 3D localization microscopy. Namely, by incorporating a physical layer in the CNN with an adjustable phase modulation, we can jointly learn the network weights and the phase mask via backpropagation. Our approach is highly flexible and can be easily adapted to any single-molecule blinking/non-blinking dataset. We validated our approach on simulations showing superior performance to existing methods. Future work will demonstrate the method on experimentally acquired data.