Interpretation of human motion in image sequences using situation graph trees

Conference Article


CVC Workshop on Computer Vision: Advances in Research & Development (162)





Doc link


Download the digital copy of the doc pdf document


Projects associated


Evaluation of human behaviour patterns in determined scenes has been a problem studied in social and cognitive sciences, but now it is raised as a challenging approach to computer science due to the complexity of data extraction and its analysis. Results obtained in this research will be helpful in cognitive sciences, above all in the human-computer interaction and the video surveillance domain. Our information source is an image sequence previously processed with pattern recognition algorithms, to extract quantitative data of the trajectories performed by the agents within the scene. Reasoning about human behavior makes necessary the inclusion of machine learning techniques, in order to represent those behaviours in a qualitative manner, allowing natural language explanation of the scene. This is achieved by means of a rule-based inference engine called F-Limette and a behaviour modelling tool based on Situation Graph Trees. The success of this approach depends on the precision of the image analysis system, the selection of suitable reasoning tools and the design of useful behaviour models. The model was tested in a street scene and the agents of interest were pedestrians. Textual descritpions are generated which qualitatively describe the observed behavior. Experimental results are provided by defining three different behaviors in a pedestrian crossing. This will allow us to confront sociological theories about human behaviour, whose quantitative base is at present being computed from statistics and not from semantic concepts.


pattern recognition.

Author keywords

Motion analysis and recognition, pattern recognition, human behavior analysis, cognitive vision

Scientific reference

P. Baiget, J. Gonzàlez, F.J. Orozco and F. Xavier Roca. Interpretation of human motion in image sequences using situation graph trees, 1st CVC Workshop on Computer Vision: Advances in Research & Development, 2006, Bellaterra, Espanya, pp. 1-6, 2006, UAB, Bellaterra, Espanya.