Building and Tracking Root Shapes

An algorithm aiming at robust and simultaneous registrations of a sequence of 3-D shapes was recently presented by Jacq et al. [IEEE Trans. Biomed. Eng., vol. 55, no. 5, 2008]. This algorithm has to carry out an implicit representation of their common root shape (RS). A particular emphasis was put on the median consensus shape, which is a specific type of RS. Unlike this previous study, mainly focusing on the algorithm foundations while dealing with very specific applications examples, this paper attempts to show the versatility of the RS concept through a set of three problems involving a wider scope of application. The first problem copes with the design of prosthetic cortical plates for the hip joint. It shows how an explicit reconstruction of the RS, coming with its consensus map, could bring out an intermediary anatomical support from which pragmatic choices could be made, thereby performing a tradeoff between morphological, surgical, and production considerations. The second problem addresses in vivo real-time shoulder biomechanics through a miniature 3-D video camera. This new protocol implicitly operates through RS tracking of the content of virtual spotlights. It is shown that the current medical-oriented protocol, while operating within expert offices through low-cost equipments, could challenge high-end professional equipments despite some limitations of the 3-D video cameras currently available. The last problem deals with respiratory motions. This is an auxiliary measurement required by some medical imaging systems that can be handled as a basic application case of the former new protocol.