Non-rigid surface matching to unmask scoliotic deformity in surface topography

A method is currently being developed for automatically modelling deformations that occur on a scoliosis patient's back over a period of time. The classic rigid matching of two measured back shapes to quantify the deformations that has occurred, suffers from the difficulty that the differences between the two backs are generally too large for matching to be effective. That is, we are typically trying to align two surfaces which are too different to be aligned. We proposed a novel non-rigid matching based on least squares approach to model the surface deformations due to scoliosis. New parameters have been investigated and incorporated into the matching algorithm. They are three scales and six shears non-rigid parameters. A 3D laser scanner is used to capture patient back scans and generated into surface topography. The surfaces are automatically matched by a least squares non-rigid matching algorithm implemented in MATLAB. Experiments on various scoliosis data demonstrate that this non-rigid matching algorithm is feasible and favourable for scoliosis deformations modelling