Inhalt

•  
  Introduction
•  
   Combining Simple Methods
•  
   Recognition Strategies
•  
   Object Context Knowledge
•  
   Proposed Integrating Framework
•  
   Outline
•  
  Image Segmentation
•  
   Basic Concepts
•  
   Image Data Driven Features and their Distances
•  
   Approaches for Segment Generation
•  
   Integrating Task Specific Knowledge to Segmentation
•  
   Model Based Top Down Segmentation
•  
   Choice of Color Image Segmentation Algorithms
•  
   Hierarchical Region Growing: Color Structure Code
•  
   Graph Based Segmentation Using the Local Variation Criterion
•  
   Feature Clustering Using Mean-Shift Algorithm
•  
   Image Segmentation by Pixel Color Classification
•  
   Perceptual Grouping of Contour Information
•  
   Summary
•  
  Object Recognition
•  
  Basic Concepts
•  
   Recognition Task: Detection, Segmentation, and Labeling
•  
   General Components of Object Recognition Systems
•  
   Object Knowledge Representation
•  
   Classifier Combination
•  
   Object Recognition Systems
•  
   Hybrid System Integrating Neural and Semantic Networks
•  
   Combining Region Based Classifiers for Recognition
•  
   Appearance Based Recognition System
•  
   Shape Based Recognition
•  
   Additional Information: Context Based Systems
•  
   Semantic Region Growing
•  
   Recognition based on Assemblage Rules
•  
   Monitoring the Assembly Construction Process
•  
   Summary
•  
   The Integrating Framework
•  
   Integrated System Architecture and Component Interaction
•  
   Common Representation of Segment and Object Information
•  
   Exemplary Effects in Data Driven Segmentation
•  
   Generating a Hierarchical Representation of Segmentation Results
•  
   Image Data Based Object Information
•  
   Summarized Characteristics of the Common Representation
•  
  Generating Hypotheses by Analyzing the Hierarchical Representation
•  
   Object Labels from Probabilistic Integration
•  
   Selecting Hypotheses for Object Regions
•  
   Information Content of the Competing Hypotheses
•  
   Additional Information Dependent on Preliminary Hypotheses
•  
   Localizing Object Information from Expectation
•  
   Integrating Additional Information
•  
  Evaluation of Competing Hypotheses
•  
   Independent Evaluation Criteria
•  
   Combination of Individual Criteria
•  
   Summary
•  
   Evaluation of Realized Integrated Systems
•  
  Realized Systems and Evaluation Conditions
•  
   Components of the Realized Integrated Recognition Systems
•  
   Test Sets and Evaluation Guidelines
•  
  The Integration of Segment Information
•  
   Evaluation Strategy
•  
   baufix Task
•  
   Office Environment
•  
   Conclusions from Integrating Segment Information
•  
  Independent Image Data Based Object Information
•  
   Individual Evaluation of Object Information for the baufix Task
•  
   Evaluating the Object Label Integration for the baufix Task
•  
   Rule Based Analysis of the Common Representation
•  
   General Rules
•  
   Improvements by Integrating Data Based Modules
•  
   Competing Object Hypotheses
•  
   Discussing Domain Specific Restrictions for the baufix Task
•  
  Additional Knowledge Integration
•  
   Semantic Region Growing
•  
   Assembly recognition process
•  
   Expectations from Monitoring the Construction Process
•  
   Shape Based Office Object Recognition
•  
   Evaluation Scheme for Competing Hypotheses
•  
   Summarizing the Evaluation
•  
   Summary and Conclusion
•  
   The General Integrating Module
•  
   Realizing Integrated Object Recognition Systems
•  
   Conclusion
•  
   The Inspection Tool for Integrated Systems
•  
  The baufix Domain
•  
   Motivation
•  
   Object Label Alphabets
•  
   Classification Error Matrix for Probabilistic Integration
•  
   Test Set Images
•  
  The Office Domain
•  
   Recognition Task and Strategy
•  
   Test Set Images
•  
   Statistical Significance of Recognition Results
•  
   Bibliography