Mesh deformation strategies in shape optimization. Application to forensic facial reconstruction Article Swipe
This thesis is devoted to the conception, the development and the analysis of mesh deformation strategies for shape modeling, processing and comparison -as shape matching and surface reconstruction- and, in a rather independent concern, for devising a robust computational method for facial reconstruction. Facial reconstruction is about the estimation of a facial shape from the sole datum of the underlying skull and is a challenging problem in anthropology and forensic science. The main contribution of the thesis is the design of a novel method for shape matching, borrowing techniques from the shape optimization context. Under the sole assumption that the two shapes share the same topology, the desired mapping is achieved as a sequence of elastic displacements by minimizing an energy functional based on a signed distance function. Several numerical examples are presented to show the efficiency of the method.Also, a novel method for generating a closed surface mesh model of an initially non-closed source mesh model is developed. The method relies on an original PDE-based mesh evolution technique. A template shape is iteratively deformed, producing a sequence of shapes that get 'closer and closer' to the source triangulation.The second part of the manuscript deals with the development of a landmark-free, fully automated method for digital facial reconstruction. Based on techniques of continuous deformation as 'morphing' and 'warping', the proposed approach is integrated with anthropological assumptions and mechanical models.
