Abstract

Introduction A widely used technique to create scenes for computer imagery is to define each object in a given local frame and then to express these frames in a given global (or world) frame. The main advantage of such a process is that it enables to move (translate, rotate, scale) the object by simply changing the coordinates of the local frame in the world frame. Another technique (much less usual) is to choose the local frame so as the object is totally enclosed in the unit cube 2 (i.e. local coordinates are within [\Gamma1; 1] 3 ). The idea here is to associate the same canonic threedimensional domain with each object, whatever its position, orientation or scale. This canonic domain will greatly simplify the application of procedural deformations or procedural solid textures (see also [BLAN95a] and [BLAN95b] for an illustration of this idea). Moreov

Citations