Issue No. 04 - July/August (2004 vol. 24)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MCG.2004.8
Christian Hofsetz , Universidade do Vale do Rio dos Sinos
Kim Ng , STMicroelectronics San Diego Laboratory
George Chen , STMicroelectronics San Diego Laboratory
Peter McGuinness , STMicroelectronics San Diego Laboratory
Nelson Max , University of California, Davis
Yang Liu , University of California, Davis
It's difficult to get accurate depth information with computer vision's 3D estimation techniques. When the estimated depths are uncertain, conventional image-based rendering algorithms fails and the view rendering quality suffers. This article demonstrates an image-based rendering algorithm that can render good quality views even when the estimated depths are uncertain. Instead of only using the depth information per pixel, this approach computes a depth uncertainty region. We show how to extract the depth from the input images and how to compute the uncertainty region. The article shows how to render depth with uncertainty by splatting 3D ellipsoidal Gaussian kernels.
Yang Liu, G. Chen, P. McGuinness, N. Max, K. Ng and C. Hofsetz, "Image-Based Rendering of Range Data with Estimated Depth Uncertainty," in IEEE Computer Graphics and Applications, vol. 24, no. , pp. 34-42, 2004.