Li, Zhe: Vision-based manipulative gesture recognition in a human-robot interaction scenario. 2008
Inhalt
- 1 Introduction
- 2 Background Statement
- 2.1 Comprehension of Gesture
- 2.2 Manipulative Gestures
- 2.2.1 Discussion on Manipulation Recognition
- 2.2.2 An Indoor Scenario for Observation
- 2.2.3 Corresponding Problems in Vision-Based Approach
- 2.3 State of Art
- 2.3.1 2-D and 3-D Gesture Recognition
- 2.3.2 State-Based and Trajectory-Based Approaches
- 2.3.3 Gesture Recognition with Context
- 2.3.4 Task Learning
- 2.4 Contributions
- 3 Feature Extraction for View-Variant Observation
- 3.1 Gesture Recognition with Different View-Angles
- 3.2 Low-Level Image Processing
- 3.3 Feature Vector Construction for Manipulations
- 3.4 Summary
- 4 Manipulative Primitive Detection
- 4.1 Related Work
- 4.2 Elementary Trajectory Recognition and Spotting
- 4.3 Manipulative Primitive Modeling and Detection
- 4.3.1 Primitive Model
- 4.3.2 Sequential Monte Carlo Method for Trajectory Matching
- 4.3.3 Particle Filter Realized Hidden Markov Model Matching
- 4.4 Evaluation
- 4.5 Summary
- 5 Manipulative Task Modeling and Recognition
- 5.1 Related Work
- 5.2 Models for Symbol Sequences
- 5.3 Layered Representation of Manipulative Task
- 5.4 Coupling of Top-Down and Bottom-Up Processes
- 5.5 Task Recognition in an Office Scenario
- 5.6 Hierarchical Representation
- 5.7 Summary
- 6 Manipulative Task Learning
- 6.1 Task Learning for Human-Robot Interaction
- 6.2 Semi-Supervised Incremental Task Learning
- 6.3 Extending Task Knowledge
- 6.4 Experiment
- 6.5 Summary
- 7 Summary and Conclusion
- Bibliography
- Index