Breadcrumb
Haar-Fisz transformations
Supervisor: Guy Nason
Theme: Multiscale Methods
Haar-Fisz transforms are a novel way in which to use modern multiscale techniques (such as wavelets) to effect variance stabilization and pull the distribution of data closer to the normal distribution. A recent innovation is the data-driven Haar-Fisz transform which stabilizes data even where the precise distribution is not known a priori. Several potential projects are available that can further extend and develop the Haar-Fisz paradigm. For example, an investigation into using Haar-Fisz transformation to enhance video shot in low-light conditions.