We present a dual-stage neural network architecture for analyzing fine shape details from microscopy recordings in 3D. The system, tested on red blood cells, uses training data from both healthy donors and patients with a congenital blood disease. Characteristic shape features are revealed from the spherical harmonics spectrum of each cell and are automatically processed to create a reproducible and unbiased shape recognition and classification for diagnostic and theragnostic use.
@article{arxiv.2005.08040,
title = {Artificial neural networks for 3D cell shape recognition from confocal images},
author = {G. Simionato and K. Hinkelmann and R. Chachanidze and P. Bianchi and E. Fermo and R. van Wijk and M. Leonetti and C. Wagner and L. Kaestner and S. Quint},
journal= {arXiv preprint arXiv:2005.08040},
year = {2020}
}