In this paper we link, through simple examples, between three basic approac
hes for signal and image denoising and segmentation: (1) PDE axiomatics, (2
) energy minimization and (3) adaptive filtering. We show the relation betw
een PDE's that are derived from a master energy functional, i.e. the Polyak
ov harmonic action, and non-linear filters of robust statistics. This relat
ion gives a simple and intuitive way of understanding geometric differentia
l filters like the Beltrami flow. The relation between PDE's and filters is
mediated through the short time kernel.