Nous avons la chance d’acceuillir Mohamed Adel du groupe de recherche « Computer Vision, Imaging and Machine Intelligence » CVI2 , Université du Luxembourg. Il nous présentera ses travaux récents « Leveraging Equivariant Features for Absolute Pose Regression » . Pour ceux qui se trouvent sur le site du Creusot, vous pouvez également assister au séminaire en Amphi B à Condorcet.
Abstract: There is a lot of excitement around the potential of deep learning for geometric tasks. Yet they are not able to compete with 3D geometry-based methods in pose estimation. In this talk, we will present our recent work on « Leveraging Equivariant Features for Absolute Pose Regression ». Where we demonstrate how a translation and rotation equivariant Convolutional Neural Network directly induces representations of camera motions into the feature space.
Bio: Mohamed Adel is a Ph.D. candidate in the CVI2 research group at SnT – University of Luxembourg, working under the supervision of Prof. Djamila Aouada, in collaboration with LMO. Mohamed received his Master’s degree in computer vision from the University of Burgundy Franche-Comte (France), in 2019. His research interests are in computer vision and deep learning focusing on pose estimation, and space situational awareness.