Statistical Shape Analysis for Computer Aided Spine Deformity Detection

In this paper we describe a medical application where we exploit surface properties (measured in form of 3D-Range scans of the human back) to derive a-priori unknown additional properties of the proband, that otherwise can only be acquired using multiple x-ray recordings or volumetric scans as CT or MRI. On the basis of 274 data sets, we perform classification using statistical shape analysis methods. Consistent parameterization and alignment is achieved on the basis of only few anatomic landmarks. As our choice of landmarks is easy to detect on the human body, our approach is feasible for screening applications that can be expected to have much impact on the early detection and later treatment of spine deformities, in particular scoliosis.