Enhanced Vertebral Segmentation and Cobb angle Calculation using Advanced Instance Segmentation Techniques
VAWKUM Transactions on Computer Sciences · 22.02.2026 · 14 (1) · 15-27
DOI: 10.21015/vtcs.v14i1.2197 ↗Kurz gefasst
Fachartikel in der Zeitschrift VAWKUM Transactions on Computer Sciences, online erschienen am 22.02.2026. Verfasst von Inzamam Ul Haq und 4 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben.
Abstract (Originalsprache)
Scoliosis, a spinal deformity affecting children and adolescents, is quantified using the Cobb angle. Traditional manual measurement by clinicians is time-consuming and subject to variability. This study introduces an automated method using the YOLACT++ instance segmentation model to detect vertebrae and calculate the Cobb angle on the SpineWeb dataset. By identifying the most tilted vertebrae using principal component analysis, our approach achieves a Symmetric Mean Absolute Percentage Error (SMAPE) of 8.11%, outperforming previous segmentation-based methods. This demonstrates improved accuracy and reliability, with potential for clinical decision support. The source code and other details are available at github link. https://github.com/inzamam-ulhaq-collab/Cobb-angle-measurement-code.git
Zitieren
APA
Haq, I. U., Ullah, I., Khalil, F. A., Jan, S., & Khan, I. A. (2026). Enhanced Vertebral Segmentation and Cobb angle Calculation using Advanced Instance Segmentation Techniques. VAWKUM Transactions on Computer Sciences, 14(1), 15-27. https://doi.org/10.21015/vtcs.v14i1.2197
Vancouver
Haq IU, Ullah I, Khalil FA, Jan S, Khan IA. Enhanced Vertebral Segmentation and Cobb angle Calculation using Advanced Instance Segmentation Techniques. VAWKUM Transactions on Computer Sciences. 2026;14(1):15-27. doi:10.21015/vtcs.v14i1.2197
BibTeX
@article{Haq2026,
author = {Haq, Inzamam Ul and Ullah, Izaz and Khalil, Fazli Amin and Jan, Sakin and Khan, Israr Ahmed},
title = {Enhanced Vertebral Segmentation and Cobb angle Calculation using Advanced Instance Segmentation Techniques},
journal = {VAWKUM Transactions on Computer Sciences},
year = {2026},
volume = {14},
number = {1},
pages = {15--27},
doi = {10.21015/vtcs.v14i1.2197}
}
EUVIMED