Experimental data from the the paper "Measuring Laser Beams with a Neural Network."

See https://github.com/Dipolar-Quantum-Gases/nn-beam-profiling for code pertaining to the dataset 
and the paper. Below we show the tree structure for the repository.

experimental_data/
├── fimgs/
├── imgs/
└── text/
    ├── data_train.json
    └── data_val.json

The repo contains contains experimental Gaussian beam data (experimental_data) which comprises
both images and the ground truth annotations associated with the images.

There is a directory for monochrome images (fimg) which are .tiff images and a directory (img)
for the RGB .png images which are mapped from the monochrome images using the Viridis colormap. 

Lastly, there is a text directory (text) which contains two .json files—one for the training 
data and one for the validation data. These contain the file paths, labels, fits, 
regions-of-interest, rotated-regions-of-interest, masks and keypoints for all the Gaussian beams. 
For a more detailed  explanation of the .json annotations see the github repository at the link 
above. 