Robust motion estimation for qualitative dynamic scene analysis / von MSc-EE. Mahmoud Ali Ahmed Mohamed ; Erster Gutachter: Prof. Dr.-Ing. Bärbel Mertsching, Zweiter Gutachter: Prof. Dr. Domenec Puig. Paderborn, 2019
Content
- Introduction
- Active Vision System
- Motion Estimation
- Problem Formal Definition
- Challenging of Optical Flow Estimation
- Contributions of This Work
- Thesis Outline
- Related Literature
- Overview
- Motion Field
- Motion Detection
- Background Subtraction (BS)
- Mixture of Gaussians (MoG)
- Kernel Density Estimator (KDE)
- Codebook Construction (CC)
- Motion Estimation for Moving Camera and Objects
- Optical Flow
- Data Conservation For Differential Optical Flow
- Brightness Constancy Assumption (BCA)
- Gradient Constancy Assumption (GCA)
- Differential Optical Flow Violation
- Differential Optical Flow Estimation
- Local Smoothness Based Methods
- Global Smoothness Based Methods
- Combined Local Global (CLG) Based Methods
- Variational Optimization Framework
- Coarse-To-Fine Optimization
- Image Texture
- Structure Texture Decomposition via Total Variation (ROF)
- Normalized Cross Correlation (NCC)
- Census Transform (CT)
- Histogram of Oriented Gradients (HOG)
- Distributed Average Gradient (DAG)
- Local Directional Pattern (LDP)
- Tracking of Multiple moving Objects
- Metrics for Accuracy and Performance
- Proposed Multi-Resolution Optimization
- Large Displacements Optical Flow Problem
- Related Work
- The Proposed Approach
- Evaluation and Experimental Results
- Middlebury Training Dataset
- Middlebury Test Dataset
- Large Displacements Optical Flow Dataset
- Real Application
- Summary
- Proposed Robust Optical Flow Estimation
- Related Work
- Texture Constancy Assumption
- Color Texture
- Experiments and Evaluation
- Synthetic Illumination Changes
- KITTI 2012 Datasets
- MPI Dataset
- Middelebury Dataset
- Evaluation of Color Texture
- Evaluation of the Execution Time
- Summary
- Proposed Monocular Epipolar Line Constraint
- Introduction
- Epipolar Constraint
- Optical Flow Model
- Enhancement of Fundamental Matrix
- Experimental Results
- Epipolar Line Constraint for Sparse Optical Flow
- Epipolar Line Constraint for Dense Optical Flow
- Fundamental Matrix Re-estimation
- Challenging Sequences
- KITTI Evaluation
- Conclusion
- Proposed Real-time Multi-Objects Tracking
- Introduction
- Related Work
- The Proposed Approach
- Motion Detection
- Motion Estimation and Multi-Object Tracking
- Occlusion Handling
- Camera Motion Stabilization
- Experimental Results
- Multi-Objects Tracking Accuracy
- Datasets
- Objects Tracking with a Mobile Robot
- Real-Time Performance
- Outdoor Scenarios
- Summary
- Summary and Outlook
- Bibliography
- List of Publications
- List of Notations
- List of Abbreviations
- List of Tables
- List of Figures