1. Scene Identification based on
Concept Learning
Beatrice van Eden
- Part time PhD Student at the University of the Witwatersrand.
- Fulltime employee of the Council for Scientific and Industrial Research.

2. Index
• Broad problem statement
Give an overview of the research
I am busy with.
• Identify items in a scene 3D
Current work on the project.
• Continued work
Highlights of some of the reading
I did.

3. Broad Problem Statement
• How can we give a mobile robot the capability to
continuously and autonomously form concepts
of its environment?

4. Identify items in a scene 3D
• Marius Muja of University of British Columbia (2009)
• Manipulating tabletop objects: one of the main tasks of a robot in
a household environment
• Setting/cleaning the table
• Serving drinks
• Usual tabletop objects are difficult (no texture, transparent)
• Techniques based
on local features fail
• RANSAC – Random
sample consensus
(voting scheme to
find optima fitting
result)

5. Identify items in a scene 3D
• Grasping of objects requires precise object localization
• Bounding box around the object is not enough
• Know the grasp
• Extended the 2D chamfer matching approach to 3D
• Chamfer – calculates distance between two images

6. Identify items in a scene 3D
• Two stage approach
• Bottom-up object localization
• Determines probable object locations
• Table plane detection and removal
• Point cloud clustering
• Top down model fitting
• Determines exact object pose and identity
• Find the model with the best correspondence to the point cloud
• ICP-like (Iterative Closest Point) algorithm

7. Identify items in a scene 3D
• tabletop objects" package
• The 3D model fitter determines
• Object identity
• Object pose
• Grasp pose
• Object mesh - used in the planning stage for a more precise collision
map
• Integrated with
the planning pipeline

8. Continued work
• Backtracks a bit.
• Go back to the technical sections of papers I already read
• Go through the pseudo code of the work in this papers
• Test one or two more things I am unsure about in the 2D object
detection I showed you last week.

9. Thank you