Automatic Spine Segmentation in CT Scans

Gabor Revy, Daniel Hadhazi, Gabor Hullam

2023

Abstract

The segmentation of the spine can be an essential step in computer-aided diagnosis. Current methods aiming to handle this problem generally employ an explicit model of some type. However, to create an adequately robust model, a high amount of properly labeled diverse data is required. This is not always accessible. In this research, we suggest an explicit model-free algorithm for spine segmentation. Our approach utilizes expert algorithms that are built on medical expert knowledge to create a spine segmentation from thoracic CT scans. Our system achieves an IoU (intersection over union) value of 0.7103±0.051 (mean±std) and a DSC (Dice similarity coefficient) of 0.8295±0.0343 on a subset of the CTSpine1K dataset.

Paper Citation

in Harvard Style

Revy G., Hadhazi D. and Hullam G. (2023). Automatic Spine Segmentation in CT Scans. In Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 2: BIOIMAGING; ISBN 978-989-758-631-6, SciTePress, pages 86-93. DOI: 10.5220/0011660000003414

in Bibtex Style

@conference{bioimaging23,
author={Gabor Revy and Daniel Hadhazi and Gabor Hullam},
title={Automatic Spine Segmentation in CT Scans},
booktitle={Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 2: BIOIMAGING},
year={2023},
pages={86-93},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011660000003414},
isbn={978-989-758-631-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 2: BIOIMAGING
TI - Automatic Spine Segmentation in CT Scans
SN - 978-989-758-631-6
AU - Revy G.
AU - Hadhazi D.
AU - Hullam G.
PY - 2023
SP - 86
EP - 93
DO - 10.5220/0011660000003414
PB - SciTePress