Hip dysplasia is the most common congenital skeletal anomaly worldwide, with an incidence of 2–4%. If left untreated, it can lead to osteoarthritis and necessitate joint replacement at an early age. In newborns, however, hip dysplasia can be fully corrected through non-surgical measures if detected early. In Germany, screening is therefore part of the statutory U-examinations, during which both hip joints are assessed by ultrasound to determine the degree of femoral head coverage by the acetabulum.

This ultrasound examination is demanding even for experienced clinicians and is therefore highly operator-dependent. Traditionally, a specific sectional view is sought, in which various hip angles can be measured according to the Graf method. The aim of the JOINT project is to simplify and enhance this examination through the use of artificial intelligence. By recording sweeps rather than individual sectional images, three-dimensional reconstruction of the ultrasound scans becomes possible. This enables precise 3D reconstruction of hip anatomy using affordable and radiation-free ultrasound diagnostics.