Statistical recognition of motion patterns

This paper addresses the problem of people detection in dynamic environments in real-time and the extraction of typical motion patterns. The need for estimation of peoples movements with a mobile robot is essential for many applications in the field of service robotics. One key prerequisite for a machine to interact intelligently with people is its ability to recognize humans and their behavior. This paper presents an approach for tracking people in dynamic environments and to determine their behavior.

Authors

• Jörg Illmann
• Boris Kluge
• Erwin Prassler

Keywords

• Pattern recognition
• Cameras
• Mobile robots
• Robustness
• Layout
• Data mining
• Image segmentation
• Robot vision systems
• Intelligent robots
• Machine intelligence
• Motion Patterns
• Dynamic Environment
• Mobile Robot
• Type Of Motion
• Environment For People
• Pattern Extraction
• Spatial Information
• Color Vision
• Color Features
• Object Tracking
• Color Information
• Illumination Changes
• Gap Distance
• Color Distribution
• Laser Ranging
• Motion Vector
• Color Histogram
• Vector Quantization
• Tracking Procedure
• Adjacent Clusters
• People Tracking
• Automated Guided Vehicles
• Tracking Mechanism
• Background Scene
• Successive Scans
• Current Scan
• Spatial Cognition
• Tracking Process
• Tracking System
• Maximum Flow