Detecting Spinal Abnormalities from Biomechanical Angles

The shape and orientation of the spine and pelvis carry strong signals about spinal health. This project uses biomechanical angle measurements as features to build a machine-learning model that flags spinal abnormalities and supports clinical decision-making.

The approach

Rather than working from images, the model works from biomechanical angles — numeric measurements describing the orientation and tilt of the spine and pelvis. These features are cleaned, scaled, and used to train a supervised classifier that distinguishes normal from abnormal cases, validated with cross-validation to keep the results honest.

Technologies used

• Python: the full modeling workflow.

• scikit-learn: supervised classification, feature scaling, cross-validation, and evaluation metrics.

• Pandas & NumPy: loading, cleaning, and transforming the biomechanical feature set.

• Matplotlib: visualizing feature distributions and model results.

Impact

The model improves diagnostic accuracy on biomechanical data and gives clinicians an additional, data-driven signal to aid decision-making.

---

Tech stack: Python, scikit-learn, Pandas, NumPy, Matplotlib, supervised classification, cross-validation.