The greatest influence to my work comes from Meyers, Lounsbery, and Schroeder and Sweldens. All these systems first perform a wavelet analysis on the data, and then create simpler representations by adding selectively wavelets.
The closest work to mine is by Certain et al. There the aim is to dynamically choose the amount of detail that is shown. A model is refined by adding more wavelet coefficients to it. Coarsening of the model is dealt with storing several "snapshots" with different amounts of detail. If a simpler representation is needed, the current model is replaced by one of these snapshots.
Our work, on the other hand, does away with these snapshots, and allows incremental removal of the wavelets the same way they are added. Our method would be also easy to add to Certain's viewer.