Experimental Data for the Paper titled "3D Structure from SMLM images using Deep Learning" The following files comprise 19 sets of 40,000 images, each set corresponding to a different rendering sigma as described in the paper. * Extracting * To extract the files, execute the following command (under Linux): cat paper_data.tar.gz.* | tar xzvf - * Organisation * The data are grouped into 19 directories, corresponding to the sigma value they were rendered at. These values are 10, 9, 8.1, 7.29, 6.56, 5.9, 5.31, 4.78, 4.3, 3.87, 3.65, 3.28, 2.95, 2.66, 2.39, 2.15, 1.94, 1.743, 1.57, 1.41 Each directory contains 40,000 FITS files - a NASA floating point image standard. The images are single channel and un-normalised. This structure is ready to be used * Recreating the data * If you have the time and compute power, you can regenerate this data set with as many or as few images as you prefer, at any sigma level. The original experimental data is available at <fill in later>. To recreate the data set you need to download the CEPRender program, available on github: https://github.com/OniDaito/CEPrender - details on how to use this program are available with the code.