System for routine surface anthropometry using reprojection registration

Range data measurement can be usefully applied to non-invasive monitoring of anthropometric changes due to disease, healing or during normal physiological processes. We have developed a computer vision system that allows routine capture of biological surface shapes and accurate measurement of anthropometric changes, using a structured light stripe triangulation system. In many applications involving relocation of soft tissue for image-guided surgery or anthropometry it is neither accurate nor practical to apply fiducial markers directly to the body. This system features a novel method of achieving subject re-registration that involves application of fiducials by a standard data projector. Calibration of this reprojector is achieved using a variation of structured lighting techniques. The method allows accurate and comparable repositioning of elastic surfaces. Tests of repositioning using the reprojector found a significant improvement in subject registration compared to an earlier method which used video overlay comparison only. It has a current application to the measurement of breast volume changes in lactating mothers, but may be extended to any application where repeatable positioning and measurement is required.