期刊名称:International Journal of Computer Technology and Applications
电子版ISSN:2229-6093
出版年度:2012
卷号:3
期号:2
页码:859-864
出版社:Technopark Publications
摘要:The reconstruction of a complex scene from multiple images is a fundamental problem in the field of computer vision. Volumetric methods of computing voxels for 2-D images captured have proven to be a strong alternative to traditional correspondence-based methods due to their flexible visibility models. Whenever any 2-D image is used to reconstruct a volume based object (voxels), the edges in the images make color consistency check fail in order to determine which voxel should be carved and which should not. This inability to detect the color variance along the edges is faced by majority of the models proposed prior. Our approach comprises additional check by evaluating the edge pixels and avoiding carving of the voxel. Moreover performing visibility check through GVC-LDI ensures a better quality of reconstruction