3-D Object Discovery Using Motion
http://www.cs.washington.edu/robotics/projects/object-discovery/In contrast to object recognition or object detection, which match data to existing object models, object discovery creates object models. Obviously we need information sources to make up for the lack of models. In this project we investigate using 3-d motion of surface patches between multiple maps of the same environment as such a cue. The output of our algorithm is a segmentation of each input 3-d scene into background and movable foreground objects.
Abstract from Sept '10 paper submission:
The performance of indoor robots that stay in a single environment can be enhanced by gathering detailed knowledge of objects that frequently occur in that environment. We use an inexpensive sensor providing dense color and depth, and fuse information from multiple sensing modalities to detect changes between two 3-D maps. We do not assume alignment between maps is given. A probabilistic model of sensor readings lets us reason about movement of surfaces. Our method makes no assumptions about object shape, large movement or surface texture. We demonstrate the ability to find whole objects in complex scenes by regularizing over surface patches.
Abstract from Sept '10 paper submission:
The performance of indoor robots that stay in a single environment can be enhanced by gathering detailed knowledge of objects that frequently occur in that environment. We use an inexpensive sensor providing dense color and depth, and fuse information from multiple sensing modalities to detect changes between two 3-D maps. We do not assume alignment between maps is given. A probabilistic model of sensor readings lets us reason about movement of surfaces. Our method makes no assumptions about object shape, large movement or surface texture. We demonstrate the ability to find whole objects in complex scenes by regularizing over surface patches.
This project is affiliated with the Robotics and State Estimation Lab.
Publications
|
Toward Object Discovery and Modeling via 3-D Scene Comparison Evan Herbst, Peter Henry, Xiaofeng Ren and Dieter Fox International Conference on Robotics and Automation, 2011. Full Paper (PDF) (Slides) |
|
RGB-D Object Discovery Via Multi-Scene Analysis Evan Herbst, Xiaofeng Ren and Dieter Fox IEEE/RSJ International Conference on Intelligent Robots and Systems, 2011. Full Paper (PDF) (Slides) |