Heat kernels of generalized laplacians-application to color image smoothing

In this paper, we explore the theory of vector bundles over Riemannian manifolds in order to smooth multivalued images. In this framework, we consider standard PDE's used in image processing as generalized heat equations, related to the geometries of the base manifold, given by its metric and the subsequent Levi-Cevita connection and of the vector bundle, given by a connection. As a consequence, the smoothing is made through a convolution with a 2D kernel, generalizing Gaussian, Beltrami and oriented kernel. In particular, we construct an extension of the oriented kernel, and illustrate it with an application to color image smoothing.

[1]  David Tschumperlé,et al.  Fast Anisotropic Smoothing of Multi-Valued Images using Curvature-Preserving PDE's , 2006, International Journal of Computer Vision.

[2]  Michèle Vergne,et al.  Heat Kernels and Dirac Operators: Grundlehren 298 , 1992 .

[3]  Ron Kimmel,et al.  A general framework for low level vision , 1998, IEEE Trans. Image Process..

[4]  Rachid Deriche,et al.  Vector-valued image regularization with PDE's: a common framework for different applications , 2003, 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings..

[5]  M. Spivak.  A comprehensive introduction to differential geometry , 1979 .

[6]  Ron Kimmel,et al.  A Short- Time Beltrami Kernel for Smoothing Images and Manifolds , 2007, IEEE Transactions on Image Processing.

[7]  Rachid Deriche,et al.  Vector-valued image regularization with PDEs: a common framework for different applications , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.