Logo image
Improved 3D-model texture mapping with region-of-interest weighting and iterative boundary-texture updating
Conference paper

Improved 3D-model texture mapping with region-of-interest weighting and iterative boundary-texture updating

Yanning Wang, Heming Sha, Weiyao Lin, Yang Hua and Chia-Wen Lin
2016 IEEE International Conference on Multimedia and Expo Workshop, ICMEW 2016, 7574763
09/2016

Abstract

Signal Processing Media Technology Computer Vision and Pattern Recognition
The target of 3D-model texture mapping is to select suitable texture contents from a set of 2D images and map them onto their corresponding 3D model for creating a 'textured' 3D object. The existing texture mapping methods still suffer from large illumination differences and the mismatch between images and 3D models. In this paper, we proposed a new approach for seamless texture mapping on 3D models. The proposed approach first introduces an region-of-interest-based weighting strategy to guide the texture-selection process, such that suitable patches can be selected from images to create a more smooth and natural texture-mapping layout. Then, an iterative update strategy is further proposed, which iteratively minimizes the boundary differences between image patches and texture-boundary fitness inside each patch, such that the patch-wise texture differences on a 3D model can be deleted. Experimental results show that our approach can effectively reduce the texture difference and achieve a more satisfactory texture-mapping result.

Metrics

1 Record Views

Details

Logo image