Published March 11, 2022 | Version v1
Preprint Open

Fast Autofocusing using Tiny Networks for Digital Holographic Microscopy

Description

This dataset contains a set of experimental holograms of a pseudo-periodic pattern ranging on 92 \({\mu}m\) with a step of ~1 \({\mu}m\) between each class of holograms. A total of 40,400 holograms are divided in 100 classes each containing a total of 400 experimental holograms. 

The experiments were carried out on a antivibration table with a DHM (by Lyncee Tec, Switzerland) equipped with a camera with a 5.86 \({\mu}m\)  pixel size (Basler acA1920-155um), a hexapode (Newport HXP50-meca) capable of precise motions along the six degrees of freedom and a micro-encoded pattern made in our clean room facility (2 x 2 \(cm^2\), period 9 \({\mu}m\), 12 bits encoding) covered with a uniform 100 nm thick aluminium layer to obtain a phase object. This pseudo periodic pattern was observed with a MO (Leica, mag 10x, NA=0.32) at wavelength \(\lambda \) = 674.99 nm.

A second file (mapping file) gives the mapping in \({\mu}m\) for each class of 400 holograms: holos_04_06_2021_hxp_dhm_100classes400Measures_positionZ.csv. 

Files

holos_04_06_2021_hxp_dhm_100classes400Measures_positionZ.txt

Files (48.7 GB)

Name Size Download all
md5:c36284bd58714e37c33cd1030d6a2cf7
48.7 GB Download
md5:aa2bae1c5c857462f9f5d4ce97b85a99
1.5 kB Preview Download