Multiscale Statistical Modeling for Poisson Signal and Image Data
Prof. Eric Kolaczyk - Department of Mathematics and Statistics, Boston University
Oct 07, 2009
In astronomy and astrophysics, often the objects we seek to monitor with satellite instruments
have important structure at multiple scales. When working with photon-counting instrumentation
in the 'low count' regime, proper handling of Poisson characteristics of the data is important.
From the mid-1990s through the early 2000's, a body of literature emerged on the multiscale
statistical modeling of Poisson signal and image data. In this talk, I will present an overview
of some of the various methods from this area. In particular, I will discuss wavelet thresholding,
multiscale probabilistic models, and multiscale piecewise polynomial models.
