Skip to Main content Skip to Navigation
Journal articles

Detection of moving foreground objects in videos with strong camera motion

Abstract : In this paper, we propose a novel method for moving foreground object extraction in sequences taken by a wearable camera, with strong motion. We use camera motion compensated frame differencing, enhanced with a novel kernel-based estimation of the probability density function of background pixels. The probability density functions are used for filtering false foreground pixels on the motion compensated difference frame. The estimation is based on a limited number of measurements; therefore, we introduce a special, spatio-temporal sample point selection and an adaptive thresholding method to deal with this challenge. Foreground objects are built with the DBSCAN algorithm from detected foreground pixels.
Document type :
Journal articles
Complete list of metadatas

Cited literature [30 references]  Display  Hide  Download
Contributor : Jenny Benois-Pineau <>
Submitted on : Tuesday, February 14, 2012 - 10:53:46 AM
Last modification on : Friday, October 23, 2020 - 4:44:04 PM
Long-term archiving on: : Tuesday, May 15, 2012 - 2:30:35 AM


Files produced by the author(s)



Daniel Szolgay, Jenny Benois-Pineau, Rémi Mégret, Yann Gaëstel, Jean-François Dartigues. Detection of moving foreground objects in videos with strong camera motion. Pattern Analysis and Applications, Springer Verlag, 2011, 14 (3), pp.311-328. ⟨10.1007/s10044-011-0221-2⟩. ⟨hal-00669915⟩



Record views


Files downloads