Normal Improvement for Point Rendering

  • Authors:
  • Thouis R. Jones;Fredo Durand;Matthias Zwicker

  • Affiliations:
  • Massachusetts Institute of Technology;Massachusetts Institute of Technology;Massachusetts Institute of Technology

  • Venue:
  • IEEE Computer Graphics and Applications
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Point models from scanned data invariably contain noise. Most denoising methodsconcentrate on positional information rather than normals, even though rendered images are affected more strongly by noise in normals than positions. This article proposes a novel method for denoising normals for point models based on the bilateral filter. The method treats the filter as a spatial deformation and updates normals iteratively. The bilateral filter is feature preserving; the extension to normals inherits this trait.