Rich and Seamless Texture Mapping to 3D Mesh Models
Texture of reconstructed models are typically recovered by mapping detailed fragments to its surface, however, the visibility of the seams are often appear at the board of patches, due to inaccuracy mesh model and registration, lighting various and surface reflections etc. To address this problem, we apply a preprocessing step to every face in the mesh model by removing its candidate fragments, which are obviously different from others. Then, selecting fragment for every face is treated as a Markov random field energy optimization (MRF) problem, consist of the proposed data and smooth terms. Finally, poisson editing is employed to adjust the color information of vertices and edges of every fragments for better color consistency among fragments. Experimental results show that our method is able to produce seamless textured models with rich color information, compared with the state of the arts.
KeywordsTexture 3D graphics and realism
We thank the editors and anonymous reviewers for their insights. Also, we thank Pengfei Wu for modifying this paper. This work is supported by NSFC under grants 61571313, funding from Sichuan Province (2014HH0048) and the Science and Technology Innovation seedling project of Sichuan (2014-046).
- 1.Baumberg, A.: Blending images for texturing 3D models. In: BMVC, vol. 3, pp. 5. Citeseer (2002)Google Scholar
- 9.Lempitsky, V., Ivanov, D.: Seamless mosaicing of image-based texture maps. In: 2007 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2007, pp. 1–6. IEEE (2007)Google Scholar
- 12.Waechter, M., Moehrle, N., Goesele, M.: Let there be color! Large-scale texturing of 3D reconstructions. In: Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T. (eds.) ECCV 2014, Part V. LNCS, vol. 8693, pp. 836–850. Springer, Heidelberg (2014)Google Scholar