1: \begin{abstract}Any procedure applied to data, and any quantity derived from 

2: data, is required to respect the nature and symmetries of the data. This

3: axiom applies to refinement procedures and multiresolution transforms as

4: well as to more basic operations like averages. This chapter discusses

5: different kinds of geometric structures like metric spaces, Riemannian

6: manifolds, and groups, and in what way we can make elementary operations

7: geometrically meaningful. A nice example of this is the Riemannian metric

8: naturally associated with the space of positive definite matrices and 

9: the intrinsic operations on positive definite matrices derived from it. 

10: We disucss averages first and then proceed to refinement operations

11: (subdivision) and multiscale transforms. In particular, we report on

12: the current knowledge as regards convergence and smoothness.

13: \end{abstract}

14: