Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revisionBoth sides next revision
teaching:gsoc2017 [2017/02/08 08:12] – [Topic 2: Realistic Grasping using Unreal Engine] liscateaching:gsoc2017 [2017/02/08 08:40] – [Topic 1: Multi-modal Cluttered Scene Analysis in Knowledge Intensive Scenarios] lisca
Line 78: Line 78:
 {{  :teaching:gsoc:topic1_rs.png?nolink&200|}} {{  :teaching:gsoc:topic1_rs.png?nolink&200|}}
  
-**Main Objective:** The main objective of this topic is to enable robots in a human environment to recognize objects in difficult and challenging scenarios. To achieve this the participant will develop software components for RoboSherlock, called annotators, that are aided by background knowledge in order to detect objects. These scenarios include stacked,occluded objects placed on shelves, objects in drawers, refrigerators, dishwashers, cupboards etc. In typical scenarios, these confined spaces also bare an underlying structure, which will be exploited, and used as background knowledge, to aid perception (e.g. stacked plates would show up as parallel lines using an +**Main Objective:** In this topic we will develop algorithms that en- 
-edge detection).+able robots in a human environment to recognize objects in diffi- 
 +cult and challenging scenarios. To achieve this the participant will 
 +develop annotators for RoboSherlock that are particularly aimed at 
 +object-hypotheses generation and merging. Generating a hypotheses 
 +essentially means to generate regions/clusters in our raw data that 
 +form a single object or object-part. In particular this entails the de- 
 +velopment of segmentation algorithms for visually challenging scenes 
 +or object properties, as the likes of transparent objects, or cluttered, 
 +occluded scenes. The addressed scenarios include stacked, occluded 
 +objects placed on shelves, objects in drawers, refrigerators, dishwash- 
 +ers, cupboards etc. In typical scenarios, these confined spaces also 
 +bare an underlying structure, which will be exploited, and used as 
 +background knowledge, to aid perception (e.g. stacked plates would 
 +show up as parallel lines using an edge detection). Specifically we 
 +would start from (but not necessarly limit ourselves to) the implemen- 
 +tation of two state-of-the-art algorithms described in recent papers: 
 +[1] Ilya Lysenkov, Victor Eruhimov, and Gary Bradski, Recognition 
 +and Pose Estimation of Rigid Transparent Objects with a Kinect Sen- 
 +sor, 2013 Robotics: Science and Systems Conference (RSS), 2013. 
 +[2] Richtsfeld A., M ̈ 
 +orwald T., Prankl J., Zillich M. and Vincze 
 +M. - Segmentation of Unknown Objects in Indoor Environments. 
 +IEEE/RSJ International Conference on Intelligent Robots and Sys- 
 +tems (IROS), 2012.
  
 **Task Difficulty:** The task is considered to be challenging, as it is still a hot research topic where general solutions do not exist. **Task Difficulty:** The task is considered to be challenging, as it is still a hot research topic where general solutions do not exist.
      
-**Requirements:** Good programming skills in C++ and basic knowledge of CMake. Experience with PCL, OpenCV is prefered. Knowledge of Prolog is a plus.+**Requirements:** Good programming skills in C++ and basic knowl- 
 +edge of CMake. Experience with PCL, OpenCV is prefered.
  
 **Expected Results:** Currently the RoboSherlock framework lacks good perception algorithms that can generate object-hypotheses in challenging scenarios(clutter and/or occlusion). The expected results are several software components based on recent advances in cluttered scene analysis that are able to successfully recognized objects in the scenarios mentioned in the objectives, or a subset of these. **Expected Results:** Currently the RoboSherlock framework lacks good perception algorithms that can generate object-hypotheses in challenging scenarios(clutter and/or occlusion). The expected results are several software components based on recent advances in cluttered scene analysis that are able to successfully recognized objects in the scenarios mentioned in the objectives, or a subset of these.
- Impressum - Datenschutz



Prof. Dr. hc. Michael Beetz PhD
Head of Institute

Contact via
Andrea Cowley
assistant to Prof. Beetz
ai-office@cs.uni-bremen.de

Discover our VRB for innovative and interactive research


Memberships and associations:


Social Media: