Texturing and Deforming Meshes with Casual Images

Abstract

Using (casual) images to texture 3D models is a common way to create realistic 3D models, which is a very important task in computer graphics. However, if the shape of the casual image does not look like the target model or the target mapping area, the textured model will become strange since the image will be distorted very much. In this paper, we present a novel texturing and deforming approach for mapping the pattern and shape of a casual image to a 3D model at the same time based on an alternating least-square approach. Through a photogrammetric method, we project the target model onto the source image according to the estimated camera model. Then, the target model is deformed according to the shape of the source image using a surface-based deformation method while minimizing the image distortion simultaneously. The processes are performed iteratively until convergence. Hence, our method can achieve texture mapping, shape deformation, and detail-preserving at once, and can obtain more reasonable texture mapped results than traditional methods.