ECVision logoCognitive Computer Vision Ontology




This is an evolving topic categorization for Cognitive Computer Vision, supported by the ECVision: European Research Network for Cognitive Computer Vision Systems. Perhaps 'ontology' is not the right word to describe it, as it isn't a hierarchical subtype tree. But it's not a glossary nor syllabus either. Perhaps it's a topic catalog? Please suggest a good descriptive noun.

People directly involved in its development are: Bob Fisher, Wolfgang Förstner, Annett Faber and Hanns-Florian Schuster.


  1. Model Learning (Survey Result)
    1. Specific approaches to learning these different types of content (See also Knowledge Representation->Content for "what" things that are learned and Recognition, Categorization and Estimation->Specific Approaches for "how" things might be recognized.)
      1. Activity/Behaviors/Processes/Dynamics
      2. Classification/Category
      3. Context/Scenes/Situations
      4. Function
      5. Objects/Parts
      6. Parameters
      7. Task Control
    2. Issues
      1. Learning Control
      2. Validation
    3. Types of Learning
      1. Case-based
      2. Reinforcement
      3. Supervised
      4. Unsupervised
  2. Knowledge Representation (Survey Result)
    1. Content (See also Model Learning->Specific Approaches for learning different types of content and Recognition, Categorization and Estimation->Specific Approaches for "how" things might be recognized.)
      1. Activity/Behavior/Processes/Dynamics
      2. Classification/Category
      3. Context/Scene/Situations
      4. Function
      5. Objects/Parts
      6. Ontologies
      7. Parameters
      8. Task Control
    2. Issues
      1. Indexing
      2. Storage
    3. Style
      1. Appearance-based
      2. Embodied
      3. Generative
      4. Geometric
      5. Logical
      6. Ontological
      7. Probabilistic
      8. Procedural
      9. Relational/Graphical
  3. Recognition, Categorization and Estimation (Survey Result)
    1. General Approaches
      1. Appearance
      2. Feature Sampling
      3. Geometric/Structural
      4. Physical Models
      5. Property
      6. Temporal (discrete or continuous)
    2. Issues
      1. Accuracy
      2. Generic Classes
      3. Labeling/Localization
    3. General Techniques
      1. Alignment
      2. Attention
      3. Search
      4. Figure/Ground
      5. Grouping/Perceptual Organization
      6. Labeling
      7. Parameter Estimation and Optimization
    4. Specific Approaches to recognizing things (See also Knowledge Representation->Content for "things" that are learned and Model Learning->Specific Approaches for "learning" different types of content).
      1. Activity/Behaviors/Processes/Dynamics
      2. Class/Category
      3. Context/Scenes/Situations
      4. Functions
      5. Objects/Parts
      6. Parameters
  4. Reasoning about Structures and Events (Survey Result)
    1. Content
      1. Appearance/Visibility
      2. Objects & Spatial structures and their organistion
      3. Tasks/Goals
      4. Events & temporal structures and their organisation
    2. Issues
      1. Performance
      2. Prediction
      3. Self-analysis
      4. Uncertainty
    3. Methods (Overview)
      1. Constraint Satisfaction
      2. Hypothesize and Verify
      3. Logical
      4. Model-based
      5. Rule-Based
      6. Statistical
  5. Visual Process Control (Survey Result)
    1. Decision Making
      1. Probabilistic
      2. Rule Based
      3. Soft Control
    2. Issues
      1. Active Sensing
      2. Goal Specification
      3. Planning
      4. Process Control & Monitoring
      5. Speed of Response
    3. Paradigms
      1. Central/Distributed
      2. Covert Control
      3. Reactive
  6. Emerging Topics (Survey Result)
    1. Vision & Language Fusion
  7. Case Studies

Comments and suggestions to: , who is really Bob Fisher.


rbf@inf.ed.ac.uk