Narrow Results

In this paper, we propose a technique for intuitive, interactive modelling of 3D shapes. The technique is based on the Level–Set Method which has the virtue of easily handling changes to the topology of the represented solid. Furthermore, this method also leads to sculpting operations that are very simple and intuitive from a user perspective. A final virtue is that the LSM makes it easy to maintain a distance field representation of the represented solid. This has a number of benefits such as simplification of the rendering scheme and the curvature computation. A number of LSM speed functions which are suitable for shape modelling are proposed. However, normally these would result in tools that would affect the entire model. To facilitate local changes to the model, we introduce a windowing scheme which constrains the LSM to affect only a small part of the model. The LSM based sculpting tools have been incorporated in our sculpting system which also includes facilities for volumetric CSG and several techniques for visualization.

Since the body shapes of industrial products contain various kinds of free-form surfaces, designers need a 3D CAD system with which they can easily design such surfaces. This paper proposes a new method of modeling the free-form surface with a NURBS boundary Gregory patch. This patch, bounded by four NURBS curves, can be easily connected to other surfaces with G¹ continuity. In this paper, two methods that modify the surface shapes are described. One method modifies the direction and length of a tangent vector on the surface boundary, and the other controls the center area shape of the surface by moving a point on the surface. We have developed these capabilities on the solid modeling toolkit called DESIGNBASE. This system enables designers to dynamically control the surface shape, as they observe the parametric lines and curvature radii.

The human eye smooths noisy data in the process of perceiving fair shape. The shape of ordered planar noisy data is determined by constructing convex regions, called gates, around each point. The data inherits its shape from the shape of the minimal length polygonal path interpolating the gates in order. Specifically, the number of inflections in the minimal length polygonal path defines the shape of the original noisy data.