作者: A. I. Audu , A. H. Sadka
DOI: 10.1109/ICCSPA.2013.6487285
关键词:
摘要: Depth image-based rendering has gained significant acceptance having been at the frontiers of understanding and realization three-dimensional free-view televisions. The appearance holes on a rendered image occurrence depth discontinuity surface object virtual which forms basis view-dependent can be significantly minimized if stringent considerations are exercised region match measures such as sum absolute differences, Sum square normalized cross-correlation used during interest point detection feature matching. This paper seeks an role played by points in generation map consequently affects quality textured viewpoint between two reference still images from existing cameras.