This dataset consists of a database of regular and distorted droplets with droplet clusters extracted from high-resolution shadowgraphy images in technical sprays. It serves as a subclass for generating synthetic training data via domain randomisation, which is subsequently used to train Deep Learning (DL) based image segmentation models.
A more detailed description is provided in the corresponding paper:
https://doi.org/10.1016/j.ijmultiphaseflow.2023.104702
Data Samples
Regular droplets:
Distorted droplets: