We address the problem of analysis of families of shapes which can be classified according to two categories: the main one corresponding usually to the coarse shape which we call the function and the more subtle one which we call the style. The style and the function both contribute to the overall shape which makes the general analysis and retrieval of such shapes more challenging. Also there is no single way of defining the style as this depends much on the context of the family of shapes used for the analysis. That is why the definition needs to be given through the examples. The straight forward way of finding the shape descriptors ’responsible’ for a given category would be to use well known statistical methods and find through them such descriptors with which we are able to classify shapes according to a given category. When a function is dominating this approach might not suffice - we might be unable to find a set descriptors which are independent of a given function. We show how to decouple the effect of the style from that of the function by considering the shapes of the same function but different styles. We also propose a metric coanalysis approach: if two styles are similar this similarity should be reflected across different functions. We show the usability of our methods first on the example of a number of chess sets which our method helps sort. Next, we investigate the problem of finding a replacement for a missing tooth given a database of teeth.
|Title of host publication||Proceedings of the MICCAI workshop on Mesh Processing in Medical Image Analysis (MeshMed)|
|Publication status||Published - 2011|
|Event||MICCAI workshop on Mesh Processing in Medical Image Analysis - Toronto, Canada|
Duration: 18 Sep 2011 → …
|Workshop||MICCAI workshop on Mesh Processing in Medical Image Analysis|
|Period||18/09/2011 → …|