Simultaneous learning of motion and appearance
Résumé
A new learning method for motion estimation of objects with significantly varying appearance is proposed. Varying object appearance is represented by a low dimensional space of appearance parameters. The appearance mapping and motion estimation method are optimized simultaneously. Appearance parameters are estimated by unsupervised learning. The method is experimentally verified by a tracking application on sequences which exhibit strong variable illumination, non-rigid deformations and self-occlusions.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...