Three-dimensional surface curvature estimation using quadric surface patches

Recent advances in 3D scanning technology have enabled the development of interesting applications of 3D human body modelling and shape analysis, especially in the areas of virtual shopping, custom clothing and sizing surveys for the clothing industry. Most of the current applications have so far been concerned with automatic tape measurement extraction, i.e. simulation of the manual procedure for extracting a set of body measurements by conventional means, such as tapes and callipers. This has been an important advance since it has made it possible to extract in a non-intrusive manner sets of measurements in a few seconds rather than in the usual 40-45 minutes that also involve intrusive physical contact. However, this approach has a number of problems, and also fails to exploit the full potential of 3D imaging technology, as the 3D sets obtained by a scanner are still reduced to a collection of 1D measurements in order to describe body shape. The approach we present is a method for detecting significant geometric features on the 3D body surface. These features (such as ridges and umbilic points) require no a-priori anatomical information and can be used for driving matching and modelling algorithms such as deformable Active Shape Models. The approach is truly hardware-independent and will work on any set of reasonably complete 3D data, whether it is a raw point cloud or a pre-processed, canonical-typ