Deep Learning-Based Automated Magnetic Resonance Image Segmentation of the Lumbar Structure and Its Adjacent Structures at the L4/5 Level

(1) Background: This study aims to develop a deep learning model based on 3D Deeplab V3+ network automatically segment multiple structures from magnetic resonance (MR) images at the L4/5 level. (2) Methods: After data preprocessing, modified of was used for automatic segmentation MR We performed five-fold cross-validation evaluate performance model. Subsequently, Dice Similarity Coefficient (DSC), precision, and recall were also assess model’s performance. Pearson’s correlation coefficient analysis Wilcoxon signed-rank test employed compare morphometric measurements reconstruction models generated by manual segmentation. (3) Results: The obtained an overall average DSC 0.886, precision 0.899, 0.881 sets. Furthermore, all morphometry-related revealed no significant difference between ground truth Strong linear relationships correlations in automated (4) Conclusions: found it feasible perform images, which would facilitate lumbar surgical evaluation establishing