3D object recognition by neural trees

Author

Foresti, G.L. ; Pieroni, G.G.

Author_Institution

Dept. of Math. & Comput. Sci., Udine Univ., Italy

Volume

3

fYear

1997

Firstpage

408

Abstract

In this paper, a two stage method for 3D object recognition from range images is presented. The first stage extracts local surface features from the input range images. These features are used in the second stage to group image pixels into different surface patches according to the six surface classes proposed by the differential geometry. A neural tree architecture whose nodes are perceptrons without hidden layers and with sigmoidal activation functions is used. A new strategy is proposed to split the training set when it is not linearly separable in order to assure the convergence of the tree learning process. This method has been successfully applied to a large number of synthetic and real images, some of which are presented in the result section

Keywords

convergence of numerical methods; differential geometry; feature extraction; image classification; image recognition; image representation; object recognition; perceptrons; trees (mathematics); 3D object recognition; convergence; differential geometry; image pixels; local surface features extraction; neural tree architecture; neural trees; perceptrons; range images; real images; sigmoidal activation functions; surface classes; surface patches; synthetic images; training set; tree learning process; two stage method; Computer science; Feature extraction; Geometry; Image segmentation; Layout; Mathematics; Neural networks; Object recognition; Pixel; Surface morphology;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1997. Proceedings., International Conference on

Conference_Location

Santa Barbara, CA

Print_ISBN

0-8186-8183-7

Type

conf

DOI

10.1109/ICIP.1997.632139

Filename

632139