Three-dimensional neural network driving self-interference digital holography enables high-fidelity, non-scanning volumetric fluorescence microscopy.

We present a deep learning driven computational approach to overcome the limitations of self-interference digital holography that is imposed by inferior axial imaging performances. We demonstrate learning by applying the prior knowledge of a sample, a 3D deep neural network model can simultaneously suppress the defocus noise and improve the spatial resolution and signal-to-noise ratio of conventional numerical back-propagation-obtained holographic reconstruction. 3D non-scanning volumetric fluorescence microscopy can be achieved, using a 2D self-interference hologram as input, without any mechanical and opto-electronic scanning and complicated system calibration. Our method offers a high spatiotemporal resolution 3D imaging approach, which can potentially benefit, for example, the visualization of dynamics of cellular structure and measurement of 3D behavior of high-speed flow field.