Abstract:
In this paper, we take advantage of extended superquadrics to represent and recognize the entire set of 36 geons. Extended superquadrics are novel volumetric shape models...Show MoreMetadata
Abstract:
In this paper, we take advantage of extended superquadrics to represent and recognize the entire set of 36 geons. Extended superquadrics are novel volumetric shape models that include superquadrics as a special case. An extended superquadric model can be deformed in any direction because it extends the exponents of the superquadric model from constants to functions of the latitude and longitude angles in the spherical coordinate system. Thirteen features derived from the extended superquadric parameters are recovered in order to distinguish between all 36 geon classes. Classification error rates are estimated for the nearest neighbor classifier and backpropagation neural network. Both simulated data (at different noise levels) and real geon models are tested in our experiments. The results are very encouraging and have significant benefits for an object recognition system.
Published in: 2002 International Conference on Pattern Recognition
Date of Conference: 11-15 August 2002
Date Added to IEEE Xplore: 10 December 2002
Print ISBN:0-7695-1695-X
Print ISSN: 1051-4651