Visualization of multidimensional shape and texture features inlaser range data using complex-valued Gabor wavelets

摘要

The paper describes a new method for visualization and analysis ofnmultivariate laser range data using complex valued non orthogonal Gabornwavelets (D. Gabor, 1946), principal component analysis and antopological mapping network. The initial data set that provides bothnshape and texture information is encoded in terms of both amplitude andnphase of a complex valued 2D image function. A set of carefully designednoriented Gabor filters performs a decomposition of the data and allowsnfor retrieving local shape and texture features. The feature vectornobtained from this method is multidimensional and in order to evaluatensimilar data features, further subspace methods to transform the datanonto visualizable attributes, such as R, G, B, have to be determined.nFor this purpose, a feature based visualization pipeline is proposednconsisting of principal component analysis, normalization and antopological mapping network. This process finally renders a R,G,Bnsubspace representation of the multidimensional feature vector. Ournmethod is primarily applied to the visual analysis of features in humannfaces but is not restricted to that