상세정보
학술대회
Synthetic Learning Set for Object Pose Estimation: Initial Experiments
- 발행일
- 201706
- 출처
- International Conference on Ubiquitous Robots and Ambient Intelligence (URAI) 2017, pp.106-108
- 협약과제
- 17PS2200, 로봇 적용 범위 확장을 위해 3종의 조인트 모듈, 최대 7자유도의 기구부 조합에 따른 제어, 인지 시스템의 자동 구성이 가능한 모듈라 매니퓰레이션 기술 개발, 김재홍
- 초록
- 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.
- 키워드
- machine learning, pose estimation, robot manipulation, Synthetic learning set
- KSP 제안 키워드
- Computer graphics, Future research directions, Object Pose Estimation, Robot manipulation, machine Learning, robotic manipulation