High-Density Surface Reconstruction of Fine Arts and Documents for Complete Reproduction and Counterfeit Detection

Surface models with unorganized triangular meshes have already been built based on the "retrieving surface topology from the ranged surface data," procedure using, for example, Delaunay tessellations from the scanned cloud points. The procedure is widely used; however, topology is ranging error sensitive. Especially, when high-density scanning is required, error problems become fatal and the procedure is not practical. To overcome the problem, a topology pre-assigned high-resolution modeling procedure is proposed. This procedure enables one to construct a complete reproduction of fine arts and documents in shape and color, to manipulate the shape of the resulting model and to make counterfeit detectable. Our detection is based on the three-layered model. The first layer is the conventional skin surface or the external color image layer. The second layer is a terrain model of the object of, for example, i in pastes of oil paintings or toner mountain ranges on the paper. The third layer is a terrain model of the basement of, for example, the canvas of an oil painting or a sheet of papyrus or paper. If the canvas is fully painted, the second layer masks the third. Only the image layer is involved in the counterfeit detection so far. In our modeling, the 3D terrain models of the second and/or third layer are additionally and mutually combined. The inter-layer relations are synergistic. We expect the multilayer models to be highly detectable in a counterfeit.