Proceedings Fourth IEEE Workshop on Applications of Computer Vision. WACV'98 (Cat. No.98EX201)
DOI: 10.1109/acv.1998.732889
|View full text |Cite
|
Sign up to set email alerts
|

Applications of omnidirectional imaging: multi-body tracking and remote reality

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Publication Types

Select...
3
1
1

Relationship

1
4

Authors

Journals

citations
Cited by 5 publications
(3 citation statements)
references
References 1 publication
0
3
0
Order By: Relevance
“…Under this point of view, our system is simpler and easy to deploy. Some earlier works [6,21,20,5] proposed approaches to monitoring activities by using ceiling mounted omnidirectional cameras [5], in particular catadioptric devices [17], that are made of a convex mirror and a camera, pointing up to the mirror. These devices are difficult to calibrate, as the shape of the mirror must be known to compensate the angle-based distortion.…”
Section: Related Workmentioning
confidence: 99%
“…Under this point of view, our system is simpler and easy to deploy. Some earlier works [6,21,20,5] proposed approaches to monitoring activities by using ceiling mounted omnidirectional cameras [5], in particular catadioptric devices [17], that are made of a convex mirror and a camera, pointing up to the mirror. These devices are difficult to calibrate, as the shape of the mirror must be known to compensate the angle-based distortion.…”
Section: Related Workmentioning
confidence: 99%
“…In an early version of LOTS [48,49], the system supported pixel updates with the effective integration (blending) factor from a ¼ 0:25 (our fastest integration) down to a ¼ 0:0000610351: Our default value is a ¼ 0:0078125: Using such small fractions could, of course, lead to numerical instability, especially when using byte-images. Using double-precision images would help, but would be significantly slower.…”
Section: Adapting Background Modelsmentioning
confidence: 99%
“…The VSAM work at Maryland [11,12] included non-parametric models for background subtraction and low-level people tracking, but all the examples were color imagery with simple lighting and large targets. Finally, our work, [13], addressed detection/tracking in omni-directional video and included analysis in very challenging situations including snipers.…”
Section: Background and Previous Researchmentioning
confidence: 99%