The University of Waterloo Statistical Image Processing (SIP) Lab is primarily run by Paul Fieguth, professor and associate dean for the University of Waterloo's Systems Design Engineering program.
The research lies in the domain of multidimensional signal processing and large-scale estimation problems in which the key challenge is a computational one. The lab is focused on continuing the development of the theory and understanding of scale recursive estimation algorithms for multiresolution stochastic processes. Such algorithms use a statistically meaningful divide-and-conquer strategy to break large estimation problems into smaller pieces, leading to vast improvements in efficiency.
These multiscale estimation algorithms are capable of extremely fast assimilation (a more sophisticated form of averaging) of data, but suffer from a few drawbacks: it is difficult to determine the appropriate multiscale model for a given application of interest, the model assumes a time-invariant model underlying the measurements, and the estimates produced by the model often possess certain multiscale artifacts. The goal of the research has been to address these limitations in the context of a specific data assimilation problem involving a time-varying system: the shape of the surface of the world's oceans. This approach has allowed us to make a contribution to a problem in oceanography while simultaneously giving us access to an extensive data set on which to test our algorithms.
- Feb. 1, 2021
Canadian mennonite article series #1: an introduction to limits
“Anyone who believes exponential growth can go on forever in a finite world is either a madman or an economist.”
Kenneth Boulding, economist
We live in a finite world, with finite soil, water, and resources. Yet we in North America live in an economic system premised on indefinite growth. A collision between the finite and the indefinite is absolutely inevitable.