An Object Oriented Graph Approach for Image Representation and Query Based on Content
Eugen Ganea, Marius Brezovan · 2010
This paper presents a new method for image representation based on object-oriented graph structure used for formalizing the process of image retrieval based on content considering the information ex-tracted from the segmentation process and the semantic interpretation of this information. We used an image domain ontology for interpreting the image contents and constructed an Object Oriented Graphs (OOG) for describing each image. The instances of the ontology together with the OOG’s corresponding to the low level features extracted from the images after segmentation image are stored in an object-oriented database. The object-oriented native query system is used for the retrieval of the images from the database. Our technique, which combines the image low-level descriptors with image domain ontologies, has a good time complexity and the experiments showed that the retrieval can be conducted with good results.