Articles | Volume 15, issue 19
Research article
14 Oct 2022
Research article |  | 14 Oct 2022

Neural network processing of holographic images

John S. Schreck, Gabrielle Gantos, Matthew Hayman, Aaron Bansemer, and David John Gagne

Data sets

Data sets used in "Neural network processing of holographic images" J. S. Schreck, G. Gantos, M. Hayman, A. Bensemer, and D. J. Gagne

Model code and software

NCAR/holodec-ml: v0.1 M. Hayman, G. Gantos, J. S. Schreck, G. Wallach, A. Bansemer, D. J. Gagne, and C. Becker

Short summary
We show promising results for a new machine-learning based paradigm for processing field-acquired cloud droplet holograms. The approach is fast, scalable, and leverages GPUs and other heterogeneous computing platforms. It combines applications of transfer and active learning by using synthetic data for training and a small set of hand-labeled data for refinement and validation. Artificial noise applied during synthetic training enables optimized models for real-world situations.