Creating an Average Shoulder Model from Three-dimensional Surface Scans

This paper describes a practical method for generating an average shoulder model from an existing database of three-dimensional (3D) whole body scans. The resulting model was then applied to design and evaluate a series of protective shoulder plates. The proposed method involves segmenting body scans using well defined landmarks and then re-sampling each shoulder surface from the chosen sample to create surface correspondence across the sample. From the established correspondence, an average surface model is created.