Efficient Generation of 3D Lumbar Spine Models: An Anatomy-Driven Approach with Limited Input Parameters

Generating anatomically plausible spinal 3D models is significant for biomedical research, allowing the study of the mechanical behavior in the spinal structures under various conditions for different populations. In this paper, we present a novel method that can automatically build 3D lumbar models from initial vertebra meshes based on a small list of configuration parameters including a maximum and minimum height of the intervertebral discs and a lumbar lordosis angle. The conducted experiments show that our method accurately captures the user-input characteristics in the generated spine models, as evidenced by an RMSE value of 0.0. Furthermore, we performed validation by comparing the intervertebral disc measurements of our created models with anatomical reference values obtained from published studies. The results of these experiments demonstrate the robustness and accuracy of our method with an RSSR value of 0.1mm for standing spine models and an RSSR value of 0.55mm for supine spine models.