Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SLD: Segmentation-Based Landmark Detection for Spinal Ligaments

Lara Blomenkamp, Ivanna Kramer, Sabine Bauer, Theresa Schöche

TL;DR

This work tackles the critical need for accurate spinal ligament landmark detection across cervical, thoracic, and lumbar regions to enable realistic biomechanical spine models. It introduces a segmentation-based landmark detection (SLD) pipeline that first performs shape segmentation of 3D vertebrae via skeletonization, then applies geometry-driven, rule-based methods to place landmarks for ALL, PLL, ITL, SSL, ISL, CL, and LF using endplate boundaries, process extremities, facet surfaces, and geodesic paths. The authors provide a ground-truth dataset with expert annotations and an open-source plugin, and demonstrate superior accuracy (MAE $0.7$ mm, RMSE $1.1$ mm) and robust generalization across regions and mesh qualities, outperforming registration-based approaches. The method enables improved, patient-specific biomechanical simulations (FE and MBS) and supports reproducibility through released data and code, with future work aimed at handling highly variable ligaments and pathological geometries.

Abstract

In biomechanical modeling, the representation of ligament attachments is crucial for a realistic simulation of the forces acting between the vertebrae. These forces are typically modeled as vectors connecting ligament landmarks on adjacent vertebrae, making precise identification of these landmarks a key requirement for constructing reliable spine models. Existing automated detection methods are either limited to specific spinal regions or lack sufficient accuracy. This work presents a novel approach for detecting spinal ligament landmarks, which first performs shape-based segmentation of 3D vertebrae and subsequently applies domain-specific rules to identify different types of attachment points. The proposed method outperforms existing approaches by achieving high accuracy and demonstrating strong generalization across all spinal regions. Validation on two independent spinal datasets from multiple patients yielded a mean absolute error (MAE) of 0.7 mm and a root mean square error (RMSE) of 1.1 mm.

SLD: Segmentation-Based Landmark Detection for Spinal Ligaments

TL;DR

This work tackles the critical need for accurate spinal ligament landmark detection across cervical, thoracic, and lumbar regions to enable realistic biomechanical spine models. It introduces a segmentation-based landmark detection (SLD) pipeline that first performs shape segmentation of 3D vertebrae via skeletonization, then applies geometry-driven, rule-based methods to place landmarks for ALL, PLL, ITL, SSL, ISL, CL, and LF using endplate boundaries, process extremities, facet surfaces, and geodesic paths. The authors provide a ground-truth dataset with expert annotations and an open-source plugin, and demonstrate superior accuracy (MAE mm, RMSE mm) and robust generalization across regions and mesh qualities, outperforming registration-based approaches. The method enables improved, patient-specific biomechanical simulations (FE and MBS) and supports reproducibility through released data and code, with future work aimed at handling highly variable ligaments and pathological geometries.

Abstract

In biomechanical modeling, the representation of ligament attachments is crucial for a realistic simulation of the forces acting between the vertebrae. These forces are typically modeled as vectors connecting ligament landmarks on adjacent vertebrae, making precise identification of these landmarks a key requirement for constructing reliable spine models. Existing automated detection methods are either limited to specific spinal regions or lack sufficient accuracy. This work presents a novel approach for detecting spinal ligament landmarks, which first performs shape-based segmentation of 3D vertebrae and subsequently applies domain-specific rules to identify different types of attachment points. The proposed method outperforms existing approaches by achieving high accuracy and demonstrating strong generalization across all spinal regions. Validation on two independent spinal datasets from multiple patients yielded a mean absolute error (MAE) of 0.7 mm and a root mean square error (RMSE) of 1.1 mm.
Paper Structure (12 sections, 3 equations, 4 figures, 2 tables)

This paper contains 12 sections, 3 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Shape Segmentation of 3D Vertebrae
  4. Detection of Spinal Ligament Landmarks
  5. ALL and PLL
  6. ITL and SSL
  7. ISL
  8. CL
  9. LF
  10. Experiments and Results
  11. Discussion and Conclusion
  12. Compliance with ethical standards

Figures (4)

  • Figure 1: Steps of the proposed SLD method for detecting spinal ligament landmarks.
  • Figure 2: Steps of the skeletonization-based method for semantic segmentation of vertebrae Blomenkamp.2025.
  • Figure 3: Detection of the spinal ligament landmarks (red) on the corresponding vertebral regions (blue).
  • Figure 4: Comparison of four landmark detection methods (our proposed SLD method, AlDhamari2019ADO, Porto2021ALP, and Kramer.2024), applied to cervical, thoracic and lumbar vertebra samples.