Saliency sequential surface organization for free-form object recognition

  • Authors:

  • Kim L. Boyer;Ravi Srikantiah;Patrick J. Flynn

  • Affiliations:

  • Signal Analysis and Machine Perception Laboratory, Department of Electrical Engineering, The Ohio State University, Columbus, Ohio;Signal Analysis and Machine Perception Laboratory, Department of Electrical Engineering, The Ohio State University, Columbus, Ohio;Department of Computer Science and Engineering, University of Notre Dame, Notre Dame, Indiana

  • Venue:
  • Computer Vision and Image Understanding
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

We introduce an efficient, robust means of obtaining reliable surface descriptions, suitable for object recognition, at multiple scales from range data. Mean and Gaussian curvatures are used to segment the surface into regions of four saliency classes, each based on curvature consistency. We evaluate curvature consistency in a robust multivoting scheme. Contiguous regions consistent in both mean and Gaussian curvature are identified as the most homogeneous, and therefore (probably) the most salient, followed by those consistent in mean curvature only, followed by those consistent in Gaussian curvature only. Segments at each level of the hierarchy are extracted in the order of size, large to small, such that the most salient features of the surface are recovered first. To demonstrate an application of the work, we present an effective recognition system for free form objects based on attributed graphs constructed from the surface segmentation.