Synthetic learning set for object pose estimation: Initial experiments

Joo-Haeng Lee,Woo-han Yun,Jaeyeon Lee,Jaehong Kim
DOI: https://doi.org/10.1109/URAI.2017.7992897
2017-06-01
Abstract:We summarize a method to generate a synthetic learning set for object pose estimation in robotic manipulation tasks. Exploiting modern computer graphics techniques, our synthetic learning set satisfies the requirements both in quantitative diversity and qualitative precision. We report the partial results of initial experiments and discuss some future research directions.
What problem does this paper attempt to address?