categories: the static [2, and the dynamic [5,[12][13][14], A procedure for 3D model construction from sparsely and 18, 23] approaches. In the static approach, object models irregularly sampled points is presented. A two-stage dynamic are built by interpolatively joining sampled points together deformation process is presented which is designed to produce or approximately fitting patches to sampled points. In the desirable mesh properties despite difficult data characteristics. dynamic method, sampled points act as attractors to pull In a first phase, a mesh of springs is snapped down to the elastic materials into the shape encoded by the points. convex hull of the data. In the second phase, a pseudo-gravity There are two major types of deformable models: the demodel is used to attract the mesh points into concave surface flating [5, 14, and the inflating types 11]. Deflating patches. This modeling technique is a new contribution to dymodels embrace sampled points inside the models at the namic modeling methods. This process reduces the undesirable beginning of model building and later shrink progressively effects of oversmoothness, local concentration, and folding that result from the sparsity and randomness of sampled data. Our down to the points, while inflating models are initially experiments show that the proposed deformation process preencompassed by sampled points and then blow up to fit serves to some extent both the shape and size uniformities of the points. Deformable models can also be classified, in the patches constituting models. Furthermore, our modeling terms of the number of parameters involved, into lumped, process fits surfaces with prominent concavities without prior distributed, or hybrid models. Lumped models [14, 15, segmentation of input data.