A Hybrid Deep Learning Framework for Automated Cobb Angle Estimation in Scoliosis Assessment

What is it about?

Scoliosis is a condition where the spine curves abnormally, and doctors measure this curve using three Cobb angles. The angles are essential for deciding whether a patient needs monitoring, bracing, or surgery. However, measuring the Cobb angles by hand is slow, labor-intensive, and can vary from doctor to doctor, since each person may draw the lines slightly differently. This paper introduces a new AI system that automatically measures the Cobb angle from standard spine X-rays. Unlike previous methods that focus on only one type of information, your approach combines two important perspectives: 1) Global view: What the entire spine looks like as a whole (overall shape and curve). 2) Local view: Exact positions of individual vertebrae.

Featured Image

Why is it important?

By combining detailed vertebra information with the overall spine shape, the method improves confidence in treatment decisions, supports earlier detection of meaningful changes in a patient’s condition, and reduces the chance of unnecessary follow-ups or missed progression. For families and clinicians, it means clearer, more dependable information without adding workload. For healthcare systems, it means scalable, reproducible measurements that maintain quality regardless of who is on duty.