Abstract
This paper presents a novel technique for parameterizing malformation of the torso in scoliosis. Scoliosis is complex 3 dimensional deformity of the spine in which the trunk distorts because of the internal spinal deformity. Thus monitoring the distortion in the external torso would be beneficial in tracking important changes and possibly predicting the underlying changes in the internal spine. The technique proposed in this paper, facilitates torso surface assessment and it consists of three stages of digitizing, parameterizing and mapping. Self-Organizing Neural Networks (SNN) is used to parameterize the torso malformation. The orientation and position of the neuron in the SNN provides detailed insight regarding significant changes in the torso model. Preliminary results are presented to further illustrate the capability of the technique