Automatic Extraction of Semantic Relationships from Images Using Ontologies and SVM Classifiers

摘要: Extracting high-level semantic concepts from low-level visual features of images is a very challenging research. Although traditional machine learning approaches just extract fragmentary information images, their performance still not satisfying. In this paper, we propose novel system that automatically extracts such as spatial relationships or natural-enemy using combination ontologies and SVM classifiers. Our consists two phases. the first phase, are mapped to intermediate-level (e.g, yellow, 45 angular stripes). And then, set these classified into relevant object tiger) by SVM-classifiers. revision module which improves accuracy classification used. second based on extracted domain ontology, deduce spatial/natural-enemy between multiple objects in an image. Finally, evaluate proposed color including about 20 concepts.