<div class="csl-bib-body">
<div class="csl-entry">Kriegler, A., Beleznai, C., & Gelautz, M. (2024, July 8). <i>Learning Geometry: Rotation Representation of Symmetric Objects</i> [Poster Presentation]. ICVSS 2024 International Computer Vision Summer School, Sicily, Italy. http://hdl.handle.net/20.500.12708/199782</div>
</div>
-
dc.identifier.uri
http://hdl.handle.net/20.500.12708/199782
-
dc.description.abstract
We propose a representation usable for 6D pose estimation of symmetric objects. In existing works, this problem is usually circumvented, via 1.) evaluating with a metric that is symmetry-insensitive (e.g. VSSD), or 2.) modifying the architecture. In contrast, our representation is both unique and continuous for given symmetries, enabling symmetry-sensitive evaluation for basic architectures. We evaluate our approach on T-LESS.
en
dc.language.iso
en
-
dc.subject
rotation representation
en
dc.subject
symmetric objects
en
dc.subject
3D object orientation
en
dc.subject
T-LESS
en
dc.subject
pose estimation
en
dc.title
Learning Geometry: Rotation Representation of Symmetric Objects
