Point sets from range scanners are noisy and do not contain additional data such as normals, which makes it difficult to create a mesh for rendering. We use the moving least-squares (MLS) approximation to get normals and differential information, and approximate the surface near each point by an MLS surface. From this we obtain a splat of a new octagonal shape which approximates the local curvature at a point and allows fast and high-quality rendering. We show several results which demonstrate the practicality and benefits of our algorithm. © Springer-Verlag Berlin Heidelberg 2006.
CITATION STYLE
Kim, S. J., & Song, C. G. (2006). Rendering of unorganized points with octagonal splats. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3992 LNCS-II, pp. 326–333). Springer Verlag. https://doi.org/10.1007/11758525_43
Mendeley helps you to discover research relevant for your work.