Optimal importance sampling for tracking in image sequences: application to point tracking
Résumé
In this paper, we propose a particle filtering technique for tracking applications in image sequences. The system we propose combines a measurement equation and a dynamic equation which both depend on the image sequence. Taking into account several possible observations, the peculiar measure model we consider is a linear combination of Gaussian laws. Such a model allows us to infer an analytic expression of the optimal importance function used in the diffusion process of the particle filter. We demonstrate the significance of this model for a point tracking application. The resulting point tracker enables coping with trajectories undergoing abrupt changes, occlusion situations, large geometric deformations and noisy sequences. Its performances are shown on real world sequences.
Domaines
Traitement des images [eess.IV]
Fichier principal
Eccv04Arnaud.pdf (392.01 Ko)
Télécharger le fichier
eccv04.jpg (40.13 Ko)
Télécharger le fichier
Origine | Fichiers produits par l'(les) auteur(s) |
---|
Format | Figure, Image |
---|
Loading...