Behavior adaptive and real-time model of integrated bottom-up and top-down visual attention / Muhammad Zaheer Aziz. 2009
Inhalt
- Introduction
- Natural Visual Attention
- Visual Attention Modeling and Its Applications
- Need of Region-Based Approach
- Formulation of the Problem
- Segmentation
- Preattentive Feature Computation
- Saliency Computation
- Inhibition of Return
- Behavior Influence
- Dynamic Scenes and Overt Attention
- Fine-Grain Top-Down Visual Search
- Contributions of this Work
- Thesis Outline
- Related Literature
- Human Vision System
- Models of Natural Visual Attention
- Feature Integration Theory
- Spotlight Model
- Guided Search
- Fine-Grain Top-Down Attention
- Inhibition and Facilitation of Return
- Computational Models of Attention
- Feature Extraction and Saliency Map Construction
- Feature Map Combination and Popout
- Models for Inhibition of Return
- Modeling of Top-Down Attention
- Visual Behaviors in Attention Models
- Segmentation
- Analysis
- Proposed Region-Based Saliency Maps
- Transformation into Region List
- Selection of Color Space
- Technique Foundation and Innovations
- Color Dependant Thresholds
- Seed classification and two phase operation
- Integrated Edge and Region Homogeneity Check
- Region Construction
- Segmentation Results
- Filtration of Useless Regions
- Size Map Construction
- Color Saliency
- Shape Based Saliency
- Top-Down Saliency Maps
- Chapter Summary
- Proposed Region-Based Attention Model
- Model Architecture
- Visual Behaviors
- Bottom-Up Map Fusion
- Top-Down Map Fusion
- Pop-out Selection
- The Saccadic Memory
- Inhibition and Facilitation of Return
- Chapter Summary
- Experiments and Results
- Evaluation
- Conclusion and Outlook
- Bibliography
- List of Abbreviations
- List of Tables
- List of Figures