Schmidt, Joachim: Vision-based posture detection and tracking for interactive scenarios. 2009
Inhalt
- Introduction
- Related Approaches for Recognizing Humans
- Person Localization
- Pose Reconstruction and Motion Tracking
- Model Acquisition, Initialization and Error Recovery
- Vision for Human Robot Interaction
- Optimization Techniques
- Optimization Problems
- Deterministic Optimization Algorithms
- Probabilistic Optimization Algorithms
- Summary
- Person Localization
- Applicability to Different Scenarios
- Person Localization System Design
- 6D Point Cloud Generation
- Generation and Tracking of Object Hypotheses
- Over-Segmentation for Motion-Attributed Clusters
- Weak Model for Object Hypotheses
- Kernel Particle Filter for Object Localization
- Summary
- Body Pose Tracking
- Human Robot Interaction Scenario
- Body Pose Tracking System Overview
- Modeling the Appearence of Humans
- Articulated 3D Body Model
- The Monocular Challenge
- Image Cues for Body Pose Tracking
- Body Pose Observation Model
- Kernel Particle Filtering for Body Pose Tracking
- Refinement of the Particle Distribution
- Extracting the Best Body Pose
- Motion Models for Body Pose Tracking
- Random Noise Propagation
- Body Model Initialization
- Automatic Initialization Procedure Overview
- Face and Hands Detection
- Integration into the Body Pose Tracking System
- Summary
- System Evaluation and Optimization
- Evaluating the Person Localization
- Evaluating the Body Pose Tracking
- Marker-Based Ground Truth
- Error Measure Definition
- Evaluating the Accuracy of the Body Pose Tracking
- Automatic Parameter Optimization for Body Pose Tracking
- Evaluating the Automatic Initialization Procedure
- Applications
- Person Localization for Scene Reconstruction
- Body Pose Tracking for Object Attention
- Object Attention System Overview
- Trajectory-Based Gesture Recognition
- Object Attention
- Evaluating the System Performance
- Hand Gesture Detection using the Body Pose Tracking
- Motionese Developmental Studies
- Outlook
- Bibliography