Human behavior understanding in image sequences requires to study human interaction with the environment, because human beings behave depending on their location. Therefore, a semantic labeling of the scenario must be performed and provided to the system. In this work we present a method to automatically extract semantic information from a scenario by means of a set of existing agent trajectories and an ontology of the possible semantic regions. Our training algorithm constructs a path database which serves to infer a conceptual model of the scenario and to predict future agent trajectories, even from different camera views.
CitationBaiget, Pau; Fernández, Carles; Amato, Ariel; Roca, F. Xavier; Gonzàlez, Jordi. "Constructing a path database for scene categorization". A: 2nd Computer Vision: Advances in Research & Development (CVCRD), Bellaterra, Espanya, 2007. s. n., 2007, p. 1-6.
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder. If you wish to make any use of the work not provided for in the law, please contact: firstname.lastname@example.org