Search Results for author:
Found 2 papers, 0 papers with code
MatchNorm: Learning-based Point Cloud Registration for 6D Object Pose Estimation in the Real World
Our two contributions are general and can be applied to many existing learning-based 3D object registration frameworks, which we illustrate by implementing them in two of them, DCP and IDAM.
What Stops Learning-based 3D Registration from Working in the Real World?
We summarise our findings into a set of guidelines and demonstrate their effectiveness by applying them to different baseline methods, DCP and IDAM.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
