Detection of moving foreground objects in videos with strong camera motion

Authors:
D. Szolgay;J. Benois-Pineau;R. Megret;Y. Gaestel;J.-F. Dartigues
Affiliations:
LABRI, UMR 5800 CNRS, University of Bordeaux, Bordeaux, France and Péter Pázmány Catholic University, Budapest, Hungary;LABRI, UMR 5800 CNRS, University of Bordeaux, Bordeaux, France;IMS, UMR 5218 CNRS, University of Bordeaux, Bordeaux, France;ISPED, INSERM U 897, University of Bordeaux, Bordeaux, France;ISPED, INSERM U 897, University of Bordeaux, Bordeaux, France
Venue:
Pattern Analysis & Applications
Year:
2011

Citing 0
Cited 3

Independent motion detection in the light of the aperture problem

ICIAR'12 Proceedings of the 9th international conference on Image Analysis and Recognition - Volume Part I
Fusion of multiple visual cues for visual saliency extraction from wearable camera settings with strong motion

ECCV'12 Proceedings of the 12th international conference on Computer Vision - Volume Part III
Directional votes of optical flow projections for independent motion detection

ICCVG'12 Proceedings of the 2012 international conference on Computer Vision and Graphics

Quantified Score

Hi-index	0.00

Visualization

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.