Performance of autofocus capability of deep convolutional neural networks in digital holographic microscopy

Autofocusing of digital holograms of microscopic objects is a challenging problem. In this paper, an application of a deep learning in autofocusing is described. Its generalisation performance is analyzed.