Normalización de momentos visuales 3-D frenta a transformaciones afinesaplicaciones a formas sintetizadas
- Cantón Garbín, Manuel
- Gálvez Lamolda, José Manuel
ISSN: 0210-8712
Year of publication: 1986
Year: 19
Issue: 2
Pages: 26-31
Type: Article
More publications in: Revista de informática y automática
Abstract
In this paper, we propose a solution based ont he use of 3D moments invariant to translations, rotations and amplifications to solve the problem of visual recognition for 3D objects. The normalization technique here presented permit to operate directly on 3D moments, instead of operating on a 3D structure. Also, the number of 3D invariants that can be obtained is not limited " a priori", as it is the case for algebraic normalization techniques. The experimental results confirm the invariancy of moments invariants , and their utility in classifying and indentifying various three dimensional objects irrespective of size or frame of reference. This moments exhibit a potential for greatly reducing the amount of data and processing needed for the three dimensional object recognition.