Content-based Three-dimensional Engineering Shape Search

  • Authors:
  • K. Lou;S. Prabhakar;K. Ramani

  • Affiliations:
  • -;-;-

  • Venue:
  • ICDE '04 Proceedings of the 20th International Conference on Data Engineering
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we discuss the design andimplementation of a prototype 3D Engineering ShapeSearch system. The system incorporates multiplefeature vectors, relevance feedback, and query byexample and browsing, flexible definition of shapesimilarity, and efficient execution through multi-dimensionalindexing and clustering. In order to offermore information for a user to determine similarity of3D engineering shape, a 3D interface that allows usersto manipulate shapes is proposed and implemented topresent the search results. The system allows users tospecify which feature vectors should be used toperform the search.The system is used to conduct extensiveexperimentation real data to test the effectiveness ofvarious feature vectors for shape - the first suchcomparison of this type. The test results show that thedescending order of the average precision of featurevectors is: principal moments, moment invariants,geometric parameters, and eigenvalues. In addition, amulti-step similarity search strategy is proposed andtested in this paper to improve the effectiveness of 3Dengineering shape search. It is shown that the multi-stepapproach is more effective than the one-shotsearch approach, when a fixed number of shapes areretrieved.