Texture mapping of flat-like 3D models

Recently, thanks to the diffusion of scanning devices and the availability of powerful 3D modeling software, as well as to the improvements in the automation of the image-based modeling approach, it is foreseen that 3D models will become more and more ubiquitous in several research fields. However, 3D models obtained by these technologies often are lacking for a suitable photorealistic appearance, due to the low quality generated texture, or to the complete absence of it. A post-processing texture mapping is often used to obtain highly realistic models, through an accurate manual alignment of the model and the related texture. In this work, we propose an automatic approach for mapping the texture on flat-like 3D models, in which the depth component is smaller compared to the object's height and width, such as in bass-relieves, coins, paintings or facades. The method relies on the evaluation from the model geometry of a depth map, used to align the texture image. The results show the effectiveness of the proposed method.

1. Introduction

3D models are being readily used in a variety of applications including mechanical engineering simulations, virtual museums and archaeological sites reproduction for Cultural Heritage, scientific visualizations, entertainment industries for movies and video-games and so on. In order to increase the visual appearance and the realism of 3D models, a texture is joint to the model itself; this texture can be a uniform or a faded color, different colors or, more frequently, an image (such as a photo) of the object. Several methods for the texturing of a 3D model have been proposed in literature. Very frequently, they suffer from the need of manually selecting a number of common points between the model and the texture, so that a DLT registration method can be applied.