OUTLINE: * Generalities on nonparametric regression * Penalized regression splines * Penalized wavelet estimation with quadratic penalties * Other penalties and model selection * Extensions
Using the AFM tip, nanotubes are caught on a raw sample then deposited on a clean surface with an absolute precision better than 500nm. A nanostructured surface made of smooth Germanium dots on flat silicon was used as deposition sample. Nanotube mechanics is probed by AFM tip induced displacement. Nanotubes are shown to be blocked by Ge dots: it is impossible to induce a controlled displacement of the nanotube over ...
Recently, multi-resolution methods based on non-nested spaces were introduced to allow the visualization and approximation of functions defined on irregular triangulations. This paper comes back to these methods and shows more precisely how the subdivision/prediction/correction scheme of ordinary wavelet-based multi-resolution analysis (MRA) is also present in that framework. As an illustration, it is demonstrated how it can be applied in two of the classical issues of MRA: compression and level-of-detail ...
Let C = ?X epsilon IR(sup n) / f(X) = 0!, n epsilon ?2,3!, where f is a polynomial function. We want to approximate C by subdividing the parameter space. Most of the usual algorithms raise two problems: data structure management, and the choice of subdivision level which respects the geometry of C. This paper gives a method based on a topological approach. In this work, we specify the local ...
Reports by Corporate Author
