Model-Based Teeth Reconstruction

摘要

In recent years, sophisticated image-based reconstruction methodsrnfor the human face have been developed. These methods capturernhighly detailed static and dynamic geometry of the whole face, orrnspecific models of face regions, such as hair, eyes or eye lids. Unfortunately,rnimage-based methods to capture the mouth cavity inrngeneral, and the teeth in particular, have received very little attention.rnThe accurate rendering of teeth, however, is crucial for the realisticrndisplay of facial expressions, and currently high quality face animationsrnresort to tooth row models created by tedious manual work. Inrndentistry, special intra-oral scanners for teeth were developed, butrnthey are invasive, expensive, cumbersome to use, and not readilyrnavailable. In this paper, we therefore present the first approach forrnnon-invasive reconstruction of an entire person-specific tooth rowrnfrom just a sparse set of photographs of the mouth region. Thernbasis of our approach is a new parametric tooth row prior learnedrnfrom high quality dental scans. A new model-based reconstructionrnapproach fits teeth to the photographs such that visible teeth arernaccurately matched and occluded teeth plausibly synthesized. Ourrnapproach seamlessly integrates into photogrammetric multi-camerarnreconstruction setups for entire faces, but also enables high qualityrnteeth modeling from normal uncalibrated photographs and evenrnshort videos captured with a mobile phone.