A knowledge-based approach to the automatic algorithm selection for 3D scene annotation

Author

Truong, Hung ; Boochs, Frank ; Habed, Adlane ; Voisin, Yvon

Author_Institution

i3mainz, Univ. of Appl. Sci. Mainz, Mainz, Germany

fYear

2012

fDate

2-5 July 2012

Firstpage

225

Lastpage

230

Abstract

In this paper we present a novel approach for 3D point cloud processing with the aim of annotating objects in a scanned scene. Our method is based on human cognition to guide the 3D processing algorithms and uses semantic knowledge to manage data and identify immediate situation-dependent objectives. In particular, we have built a system that allows an automatic and flexible selection of algorithms. The selection strategy exploits knowledge to identify the geometrical features to be detected as well as the objects to be annotated at each stage of the 3D processing of the point cloud.

Keywords

feature extraction; geometry; knowledge based systems; object detection; object recognition; solid modelling; 3D point cloud processing; 3D processing algorithm; 3D scene annotation; automatic algorithm selection; data management; flexible algorithm selection; geometrical feature detection; human cognition; knowledge-based approach; object annotation; object detection; object recognition; scanned scene; selection strategy; semantic knowledge; situation-dependent objective; Algorithm design and analysis; Feature extraction; Object detection; Object recognition; Ontologies; Rail transportation; Semantics; detection algorithm; hybrid intelligent system; knowledge based systems; measurement by laser beam; object recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Science, Signal Processing and their Applications (ISSPA), 2012 11th International Conference on

Conference_Location

Montreal, QC

Print_ISBN

978-1-4673-0381-1

Electronic_ISBN

978-1-4673-0380-4

Type

conf

DOI

10.1109/ISSPA.2012.6310550

Filename

6310550