T. Arbel and F. P. Ferrie. Informative views and sequential recognition. In European Conference of Computer Vision, pages 469--81, 1996. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Reinforcement Learning and Visual Object Recognition - Paletta (Correct)
.... What is the most related work Learning to recognize objects has already been outlined in the broad framework of aspects representation [11] Sequential recognition minimizing perceptual entropy measures for the special case of three dimensional, predefined geometric shape models is described in [6, 2] in the framework of active recognition. Reinforcement learning was used to find optimal saccade sequences in 2 D object recognition [5] No work has been done so far, to the knowledge of the author, in (1) optimal recognition of (2) arbitrary 3 D objects (3) from appearance. 4. Who benefits from ....
T. Arbel and F. P. Ferrie. Informative views and sequential recognition. In European Conference of Computer Vision, pages 469--81, 1996.
Combining geometric and probabilistic structure for active.. - Herbin (1998) (Correct)
....for the specific selection of useful features, and a dynamic procedure to combine them. When a hypothesis is rejected, the system changes its viewpoint or extracts another set of features. The accumulation of pieces of evidence is obtained by restricted the set of potentially observable objects. [1] proposes an evaluation of 0 10 20 30 40 50 60 70 80 90 100 10 3 10 2 10 1 10 civil militar mixte Longueur de trajectoire Log erreur de reconnaissance Log erreur de reconnaissance pour les trois probl mes Figure 6: Logarithmic error for the couples of civil (solid line) militar ....
T. Arbel and F. Ferrie. Informative views and sequential recognition. In Proc. Fourth European Conference on Computer Vision, Cambridge, England, April 14-18, pages I:469--481. 1996.
Reinforcement Learning for Autonomous Three-Dimensional.. - Paletta, Prantl, Pinz (1998) (Correct)
.... measures is considered in the framework of purposive and animate vision [2, 5] Most applications still require exhaustive search over future states, or restricted analysis of immediate consequences [14, 27] Optimal sequential recognition has already been outlined for specific 3 D shape models [3, 11] and 2 D objects [6] whereas the evidence from the temporal order of multiple 2 D views has been integrated for classification in [30, 7, 26] The contribution of this paper is to introduce reinforcement learning methods [15] for active discrimination of arbitrary 3 D objects, with computational ....
T. Arbel and F. P. Ferrie. Informative views and sequential recognition. In European Conference of Computer Vision, pages 469--81, 1996.
Graphes d'aspects probabilis es Probabilistic Aspect Graphs - Herbin (Correct)
....hypothese est rejetee, le syst eme recherche un nouveau point de vue, sensoriel par mouvement de la camera, ou representatif dans la selection de nouvelles caracteristiques visuelles. L accumulation d evidence se produit par restriction incrementale du nombre d objets potentiellement observables. [1] propose une evaluation du caractere informatif des vues par calcul d une vraisemblance conditionnee par un modele parametrique d objet. Les vraisemblances peuvent etre incr ementalement mises a jour par une combinaison de probabilit es. 6 Conclusion et discussion Les grandeurs v ff et v fi ....
T. Arbel and F. Ferrie. Informative views and sequential recognition. In Proc. FourthEuropeanConferenceon Computer Vision, Cambridge, England, April 14-18, pages I:469-- 481. 1996.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC