Princeton > CS Dept > PIXL > Graphics > Publications Local Access 

Non-Rigid Range-Scan Alignment Using Thin-Plate Splines
Symposium on 3D Data Processing, Visualization, and Transmission, September 2004

Benedict Brown, Szymon Rusinkiewicz


At left, a merged 3D model of scans aligned using rigid-body ICP. Note the artifacts on the hair and ear due to the presence of deformation caused by calibration error. At right, the result of using our nonrigid alignment procedure.

Abstract

We present a non-rigid alignment algorithm for aligning high-resolution range data in the presence of low-frequency deformations, such as those caused by scanner calibration error. Traditional iterative closest points (ICP) algorithms, which rely on rigid-body alignment, fail in these cases because the error appears as a non-rigid warp in the data. Our algorithm combines the robustness and efficiency of ICP with the expressiveness of thin-plate splines to align high-resolution scanned data accurately, such as scans from the Digital Michelangelo Project. This application is distinguished from previous uses of the thin-plate spline by the fact that the resolution and size of warping are several orders of magnitude smaller than the extent of the mesh, thus requiring especially precise feature correspondence.

Citation (BibTeX)

Benedict Brown and Szymon Rusinkiewicz. Non-Rigid Range-Scan Alignment Using Thin-Plate Splines. Symposium on 3D Data Processing, Visualization, and Transmission, September 2004.

Paper
  PDF File

Talk
  3DPVT 2004 talk as PDF