How to Box Your Cells: An Introduction to Box Supervision for 2.5D Cell Instance Segmentation and a Study of Applications

Fabian Schmeisser, Fabian Schmeisser, Maria Caroprese, Maria Caroprese, Gillian Lovell, Gillian Lovell, Andreas Dengel, Andreas Dengel, Sheraz Ahmed

2025

Abstract

Cell segmentation in volumetric microscopic images is a fundamental step towards automating the analysis of life-like representations of complex specimens. As the performance of current Deep Learning algorithms is held back by the lack of accurately annotated ground truth, a pipeline is proposed that produces accurate 3D cell instance segmentation masks solely from slice-wise bounding box annotations. In an effort to further reduce the time requirements for the annotation process, a study is conducted on how to effectively reduce the size of the training set. To this end, three slice-reduction strategies are suggested and evaluated in combination with bounding box supervision. We find that as low as 1% of weakly labeled training data suffices to produce accurate results, and that predictions produced by a 10 times smaller dataset are of equal quality to when the full dataset is exploited for training.

Paper Citation

in Harvard Style

Schmeisser F., Caroprese M., Lovell G., Dengel A. and Ahmed S. (2025). How to Box Your Cells: An Introduction to Box Supervision for 2.5D Cell Instance Segmentation and a Study of Applications. In Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART; ISBN 978-989-758-737-5, SciTePress, pages 853-860. DOI: 10.5220/0013189800003890

in Bibtex Style

@conference{icaart25,
author={Fabian Schmeisser and Maria Caroprese and Gillian Lovell and Andreas Dengel and Sheraz Ahmed},
title={How to Box Your Cells: An Introduction to Box Supervision for 2.5D Cell Instance Segmentation and a Study of Applications},
booktitle={Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART},
year={2025},
pages={853-860},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013189800003890},
isbn={978-989-758-737-5},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART
TI - How to Box Your Cells: An Introduction to Box Supervision for 2.5D Cell Instance Segmentation and a Study of Applications
SN - 978-989-758-737-5
AU - Schmeisser F.
AU - Caroprese M.
AU - Lovell G.
AU - Dengel A.
AU - Ahmed S.
PY - 2025
SP - 853
EP - 860
DO - 10.5220/0013189800003890
PB - SciTePress