Multi-level background initialization using Hidden Markov Models
IWVS '03 First ACM SIGMM international workshop on Video surveillance
Exemplar-based background model initialization
Proceedings of the third ACM international workshop on Video surveillance & sensor networks
Unsupervised scene analysis: a hidden Markov model approach
Computer Vision and Image Understanding
Real-time object tracking with relevance feedback
Proceedings of the 6th ACM international conference on Image and video retrieval
The input-interface of Webcam applied in 3D virtual reality systems
Computers & Education
Unsupervised scene analysis: A hidden Markov model approach
Computer Vision and Image Understanding
Real-time foreground-background segmentation using codebook model
Real-Time Imaging
Robust foreground extraction technique using Gaussian family model and multiple thresholds
ACCV'07 Proceedings of the 8th Asian conference on Computer vision - Volume Part I
Base selection in estimating sparse foreground in video
ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
Background subtraction for automated multisensor surveillance: a comprehensive review
EURASIP Journal on Advances in Signal Processing - Special issue on advanced image processing for defense and security applications
| Hi-index | 0.00 |
In this paper a new probabilistic method for backgroundmodelling is proposed, aimed at the application in videosurveillance tasks using a monitoring static camera. Recently,methods employing Time-Adaptive, Per Pixel, Mixtureof Gaussians (TAPPMOG) modelling have becomepopular due to their intrinsic appealing properties. Nevertheless,they are not able per se to monitor global changesin the scene, because they model the background as a setof independent pixel processes. In this paper, we proposeto integrate this kind of pixel-based information with higherlevel region-based information, that permits to manage alsosudden changes of the background. These pixel- and region-basedmodules are naturally and effectively embedded in aprobabilistic Bayesian framework called particle filtering,that allows a multi-object tracking. Experimental comparisonwith a classic pixel-based approach reveals that theproposed method is really effective in recovering from situationsof sudden global illumination changes of the background,as well as limited non-uniform changes of the sceneillumination.