DocumentCode :
433034
Title :
Extracting projective invariant of 3D line set from a single view
Author :
Chen, Zhe
Author_Institution :
Sch. of Comput. Sci., Northwestern Polytech. Univ., Xi´´an, China
Volume :
3
fYear :
2004
fDate :
24-27 Oct. 2004
Firstpage :
2119
Abstract :
Although projective invariants do not exist for general 3D objects in a single view, they do exist for some special 3D constrained structures. So far, several kinds of structure and their invariants have been proposed, such as six points on two adjacent planes, six lines on three planes and so on. But the sorts of them are still limited, which constrains their wide use to 3D object recognition. In this paper, a novel non-coplanar line structure and its projective invariant were derived. This kind of structure can be easily found in the objects formed by intersected polyhedrons. The experimental results showed that the values of the proposed invariant remain stable in the perspective projection between 3D rigid object and its perspective images, and are slightly different over a range of viewpoints.
Keywords :
feature extraction; image recognition; object detection; 3D constrained structure; 3D line set; 3D object recognition; intersected polyhedron; noncoplanar line structure; projective invariant extraction; single view; Cameras; Computer science; Indexing; Lenses; Object recognition; Radio access networks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image Processing, 2004. ICIP '04. 2004 International Conference on
ISSN :
1522-4880
Print_ISBN :
0-7803-8554-3
Type :
conf
DOI :
10.1109/ICIP.2004.1421504
Filename :
1421504
Link To Document :
بازگشت