Static Multispectral Textures

摘要: The general surface reflectance function, whether static or dynamic, is too complex a function of 16 even 17 variables (dynamic GRF) to be entirely represented by any existing elementary mathematical model. One possible solution factorize the representation original GRF measurement space into set lower-dimensional and restricted representations. Approximating one with several partial local representations subsequently enables us characterize these factors using still non-trivial but simpler models tractable dimensionality. This chapter adaptive-modeling approaches, for which model learning from required target texture known. Such modeling are treated in form multispectral textures.