Höhere Bilddeutung 
(Knowledge-based Scene Interpretation)

Sommersemester 2008

Bernd Neumann
neumann@informatik.uni-hamburg.de

 

Aktuelles

Inhalt

Literatur

Folien

         

Inhalt der Vorlesung

Lecture 1:    Introduction
    Contents overview, motivation, aims, problem areas
Lecture 2:    Early work on scene interpretation
    Badler, Tsotsos, Hogg, Nagel, Neumann
Lecture 3:    Basic knowledge representation formalisms
    Semantic Networks, Frames, Constraints, Relational Structures
Lecture 4:    Conceptual units for scene interpretation
    Aggregates, situation trees, scenarios
Lecture 5:    Interface to low-level vision
    Primitive symbols, grounding
Lecture 6:    Modelling spatial and temporal relations
    Fuzzy predicates, Allen, RCC8, constraints
Lecture 7:    Interpretation procedures
Lecture 8:    Logical framework
    Model construction, Decription Logics       
Lecture 9:    Scene interpretation as configuration
    Stepwise construction, SCENIC
Lecture 10:    Probabilistic Guidance
    Hierarchical Bayesian Networks
Lecture 11:    Mobile Robot Localisation
    Simultaneous Localisation and Mapping (SLAM)
Lecture 12:    Case study
    Real-time scenario recognition (Orion/INRIA)
Lecture 13:    Case study
    Video-based Event Recognition
Lecture 14:    Summary and outlook

 
 
Literatur

Image Processing, Analysis and Machine Vision
M. Sonka, V. Hlavac, R. Boyle
Chapman & Hall 1993

Computer Vision
D.H. Ballard, C.M. Brown
Prentice-Hall 1982

Folien zur Vorlesung
Lecture 1: Introduction
Lecture 2: Early Work on Scene Interpretation
Lecture 3: Basic Knowledge Representation Formalisms
Lecture 4: Conceptual Units for Scene Interpretation
Lecture 5: Constraints
Lecture 6: Logics of Scene Interpretation
Lecture 7: Description Logics for Scene Interpretation
Lecture 8: Scene Interpretation as Configuration
Lecture 9: Probabilistic Models for Scene Interpretation
Lecture 10: Probabilistic Aggregate Models
Lecture 11: Mobile Robot Localisation