Identifying Foreground from Multiple Images

Wonwoo Lee 1 Woo Wontack 1 Edmond Boyer 2
2 PERCEPTION - Interpretation and Modelling of Images and Videos
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : In this paper, we present a novel foreground extraction method that automatically identifies image regions corresponding to a common space region seen from multiple cameras. We assume that background regions present some color coherence in each image and we exploit the spatial consistency constraint that several image projections of the same space region must satisfy. Integrating both color and spatial consistency constraints allows to fully automatically segment foreground and background regions in multiple images. In contrast to standard background subtraction approaches, the proposed approach does not require any a priori knowledge on the background nor user interactions. We demonstrate the effectiveness of the method for multiple camera setups with experimental results on standard real data sets.
Document type :
Conference papers
Complete list of metadatas

Cited literature [11 references]  Display  Hide  Download

https://hal.inria.fr/inria-00590236
Contributor : Team Perception <>
Submitted on : Tuesday, May 3, 2011 - 9:47:36 AM
Last modification on : Wednesday, April 11, 2018 - 1:59:02 AM
Long-term archiving on : Thursday, August 4, 2011 - 3:07:29 AM

File

wonwoo_accv07.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Wonwoo Lee, Woo Wontack, Edmond Boyer. Identifying Foreground from Multiple Images. ACCV 2007 - 8th Asian Conference on Computer Vision, Nov 2007, Tokyo, Japan. pp.580-589, ⟨10.1007/978-3-540-76390-1_57⟩. ⟨inria-00590236⟩

Share

Metrics

Record views

221

Files downloads

406