Qixing Huang: Extreme Relative Pose Estimation via Scene Completion
Автор: SIYUAN HUANG
Загружено: 2021-06-25
Просмотров: 369
Описание: Estimating the relative rigid pose between two RGB-D scans of the same underlying environment is a fundamental problem in computer vision, robotics, and computer graphics. Most existing approaches allow only limited relative pose changes since they require considerable overlap between the input scans. This talk discusses techniques extending the scope to extreme relative poses, with little or no overlap between the input scans. The key idea is to infer complete scene information about the underlying environment and match the completed scans. We discuss suitable representations for scene completion and how to integrate relative pose estimation and scene completion. Experimental results on benchmark datasets show that our approach leads to considerable improvements over state-of-the-art methods for relative pose estimation. In particular, our approach provides encouraging relative pose estimates even between non-overlapping scans.
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
