Skip to Main content Skip to Navigation
Conference papers

Quelle image met le mieux en valeur un modèle 3D ?

Marie Pelissier-Combescure 1 Géraldine Morin 1 S Chambon 1 
1 IRIT-REVA - Real Expression Artificial Life
IRIT - Institut de recherche en informatique de Toulouse
Abstract : Given an object present in an image, our purpose is to quantify how well this object is represented in this view. To do this, we define a highlighting score that allows us to rank a set of images : from the one that best showcases the object to the worst one. To quantify the showcase of an object in the image, we combine three complementary criteria into a highlighting score : its dominance, its size and the quantity of its characteristic information based on a score given by a curvilinear saliency detector. As an alternative, we consider the confidence scores of state of the art detection and classification neural networks. In order to validate the proposed approaches based on these scores, we provide a validation protocol based on a set images we generate to provide a reference classification. Our experimental results demonstrate the efficiency of our method and help understanding the behaviour of the networks. We also illustrate the interest of the approach with visual qualitative results on a real dataset.
Complete list of metadata
Contributor : Marie Pelissier-Combescure Connect in order to contact the contributor
Submitted on : Wednesday, July 6, 2022 - 11:25:22 AM
Last modification on : Wednesday, July 13, 2022 - 3:56:36 AM


Files produced by the author(s)


  • HAL Id : hal-03715237, version 1


Marie Pelissier-Combescure, Géraldine Morin, S Chambon. Quelle image met le mieux en valeur un modèle 3D ?. Congrès Reconnaissance des Formes, Image, Apprentissage et Perception (RFIAP 2022), AFRIF (Association Française pour la Reconnaissance et l'Interprétation des Formes), Jul 2022, Vannes, France. ⟨hal-03715237⟩



Record views


Files downloads