E
Emre Dogan
Researcher at University of Lyon
Publications - 7
Citations - 99
Emre Dogan is an academic researcher from University of Lyon. The author has contributed to research in topics: Pose & Strong prior. The author has an hindex of 4, co-authored 7 publications receiving 86 citations. Previous affiliations of Emre Dogan include Galatasaray University & Institut national des sciences Appliquées de Lyon.
Papers
More filters
Journal ArticleDOI
Evaluation of video activity localizations integrating quality and quantity measurements
Christian Wolf,Christian Wolf,Eric Lombardi,Eric Lombardi,Julien Mille,Julien Mille,Oya Celiktutan,Oya Celiktutan,Oya Celiktutan,Mingyuan Jiu,Mingyuan Jiu,Emre Dogan,Emre Dogan,Gonen Eren,Moez Baccouche,Moez Baccouche,Emmanuel Dellandréa,Emmanuel Dellandréa,Charles-Edmond Bichot,Charles-Edmond Bichot,Christophe Garcia,Christophe Garcia,Bulent Sankur +22 more
TL;DR: A new performance metric addressing and unifying the qualitative and quantitative aspects of the performance measures is proposed, which has been tested on several activity recognition algorithms participating in the ICPR 2012 HARL competition.
Journal ArticleDOI
Multi-view pose estimation with mixtures-of-parts and adaptive viewpoint selection
Emre Dogan,Emre Dogan,Gonen Eren,Christian Wolf,Christian Wolf,Eric Lombardi,Atilla Baskurt,Atilla Baskurt +7 more
TL;DR: A new method for human pose estimation which leverages information from multiple views to impose a strong prior on articulated pose and significantly decreases the estimation error compared to single-view results is proposed.
Proceedings ArticleDOI
Activity recognition with volume motion templates and histograms of 3D gradients
TL;DR: Experiments show that the new method for activity recognition based on a view independent representation of human motion outperforms the original HoG3D method.
Dissertation
Estimation de pose humaine et reconnaissance d’action par un système multi-robots
TL;DR: Ainsi, nous avons propose un modele de flexible mixtures-of-parts multi-vues inspire par la methodologie classique de structure pictural, qui verifie that l'utilisation of representations independantes of the vue et l'integration d'informations a partir of points de vue multiples ameliore the performance pour les tâches ciblees.
Dissertation
Human pose estimation and action recognition by multi-robot systems
TL;DR: In this paper, the authors introduce a representation spatio-temporelle independante du point de vue, which can capture le mouvement de la personne en utilisant un capteur de profondeur and encode en 3D for le representer.