3D Trajectory Extraction from 2D Videos for Human Activity Analysis.

Long description: The present dissertation addresses the problem of extracting 3D trajectories of objects from 2D videos. The reason of this is the theory that these trajectories symbolise high-level interpretations of human activities. A 3D trajectory of an object means its sequential positions in...

Descripción completa

Detalles Bibliográficos
Clasificación:Libro Electrónico
Autor principal: Boukhers, Zeyd
Formato: Electrónico eBook
Idioma:Inglés
Publicado: Berlin : Logos Verlag Berlin, 2017.
Colección:Studien Zur Mustererkennung Ser.
Temas:
Image processing > Digital techniques.
Traitement d'images > Techniques numériques.
digital imaging.
Image processing > Digital techniques
Acceso en línea:Texto completo
Tabla de Contenidos:
  • Intro
  • 1 Introduction
  • 1.1 Problem
  • 1.2 Contribution
  • 1.3 Overview
  • 2 RelatedWork
  • 2.1 Object Detection
  • 2.1.1 Segmentation
  • 2.1.2 Exhaustive Search
  • 2.1.3 Superficial Search
  • 2.2 Trajectory Extraction
  • 2.2.1 3D Object Tracking
  • 2.2.2 3D Trajectory Extraction from Surveillance Videos
  • 2.2.3 3D Trajectory Extraction from Moving Cameras
  • 2.3 Camera Parameters Estimation
  • 2.3.1 Focal Length Estimation
  • 2.3.2 Camera Odometry Estimation
  • 2.4 Trajectory Analysis
  • 2.4.1 Convoy Detection
  • 2.4.2 Suspicious Activity Detection
  • 3 Object Detection
  • 3.1 Object Detection in Surveillance Videos
  • 3.1.1 Clustering for One Frame
  • 3.1.2 Clustering Improvement for a Frame Sequence
  • 3.2 Object Detection in Still Images
  • 3.2.1 Systematic Search
  • 3.2.2 Classification
  • 3.3 Summary
  • 4 3D Trajectory Extraction from Surveillance Videos
  • 4.1 Initialisation
  • 4.2 Best Frame Selection
  • 4.3 Particle Filter
  • 4.3.1 Transition Distribution
  • 4.3.2 Weight Computation
  • 4.4 Coordinate Conversion
  • 4.5 Summary
  • 5 3D Trajectory Extraction from Moving Cameras
  • 5.1 Focal Length Estimation
  • 5.2 Object's Depth Estimation
  • 5.3 Observation Model
  • 5.3.1 Camera Observation
  • 5.3.2 Object Observation
  • 5.4 Motion Model
  • 5.4.1 Camera Motion
  • 5.4.2 Object Motion
  • 5.5 Sampling
  • 5.5.1 Object Birth
  • 5.5.2 Object Death
  • 5.5.3 Object Stay
  • 5.5.4 Object Exit
  • 5.5.5 Object Update
  • 5.5.6 Camera Update
  • 5.6 Summary
  • 6 Application: Trajectory Analysis
  • 6.1 Convoy Detection
  • 6.1.1 Noncontinuous Convoy:
  • 6.1.2 Clustering and Intersection:
  • 6.1.3 Candidate Expiring Mechanism:
  • 6.2 Suspicious Activity Detection
  • 6.3 Summary
  • 7 Experiments and Results
  • 7.1 Object Detection
  • 7.1.1 Object Detection in Surveillance Videos
  • 7.1.2 Object Detection in Still Image
  • 7.2 3D Trajectory Extraction.
  • 7.2.1 3D Trajectory Extraction from Surveillance Videos
  • 7.2.2 3D Trajectory Extraction from Moving Cameras
  • 7.3 Camera Parameters Estimation
  • 7.3.1 Focal Length Estimation
  • 7.3.2 Camera Odometry Estimation
  • 7.4 Trajectory Analysis
  • 7.4.1 Convoy Detection
  • 7.4.2 Suspicious Activity Detection
  • 8 Conclusion
  • 8.1 Summary
  • 8.2 FutureWork
  • Abbreviations
  • Mathematical Symbols
  • List of Figures
  • List of Tables
  • List of Algorithms
  • Bibliography
  • Own Publications
  • Curriculum Vitae.

Ejemplares similares