An algorithm for surface reconstruction from a polyhedron with arbitrary topology consisting of triangular faces is presented. The first variant of the algorithm constructs a curve network consisting of cubic Bézier curves meeting with tangent plane continuity at the vertices. This curve network is extended to a smooth surface by replacing each of the networks facets with a split patch consisting of three triangular Bézier patches. The remaining degrees of freedom of the curve network and the split patches are determined by minimizing a quadratic functional. This optimization process works either for the curve network and the split patches separately or in one simultaneous step. The second variant of our algorithm is based on the construction of an optimized curve network with higher continuity. Examples demonstrate the quality of the different met hods.The constructed curve network is filled with triangular Bézier patches interpolating the curve network, such that neighboring patches meet smoothly.
C-470A. Kolb et al. / Fair Surface Reconstruction using Quadratic Functions © Eurographics Association, 1995 Lounsbery et al [10], Mann [11] and Peters [15] give an overview on methods that are based on these construction steps. Basically, the existing methods can be divided into heuristic methods and optimizing methods.The main characteristic of the heuristic methods is that all free parameters of the curves and patches that are not determined by any of the constraints (interpolation conditions/smoothness conditions at vertices and along patch-boundaries) are set by heuristic rules. The resulting algorithms are, in general, very easy to implement. Furthermore they are relatively fast. The major drawback of the heuristic methods is that the resulting surfaces are often of very poor quality (see Mann [11]).On the other hand, optimizing methods try to overcome the shortcomings of the traditional heuristic methods utilizing the idea of variational design. Nielson [14] proposes a method applying the basic ideas of the minimum norm networks known from functional scattered data interpolation (see [13] and Section 2.2) to the general situation. Nielson uses a quadratic functional to optimize a curve network. This reduces the optimization process to solving a linear system (see Section 2.1). The curve network is filled with so-called transfinite interpolants, a very special class of rational surfaces not supported by any CAD system. The underlying representation of the curve network yields a relatively large number of unknowns for the optimization. Moreover, this representation can not be extended to curve networks with higher continuity. These seem to be the major disadvantages of this method.A . Kolb et al. /Fair Surface Reconstruction using Quadratic Functions © Eurographics Association, 1995 The representation of our curve network is based on the idea of a functional MNN (see Section 2.2). Due to the unrestricted topology of our polyhedron we can not, in general, parameterize our curve network globally. The intr...