Multi-path Learning for Object Pose Estimation Across Domains

We introduce a scalable approach for object pose estimation trained on simulated RGB views of multiple 3D models together. We learn an encoding of object views that does not only describe an implicit orientation of all objects seen during training, but can also relate views of untrained objects. Our...

Descripción completa

Detalles Bibliográficos

Autores principales:	Sundermeyer, Martin, Durner, Maximilian, Puang, En Yen, Marton, Zoltan-Csaba, Vaskevicius, Narunas, Arras, Kai O., Triebel, Rudolph
Formato:	Artículo
Publicado:	2019
Materias:	Computer Vision and Pattern Recognition Machine Learning
Acceso en línea:	https://arxiv.org/abs/1908.00151 http://arxiv.org/abs/1908.00151
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!