【摘要】 提出一种新的图像本体标注的框架,结合领域本体中概念的关系,通过层次概率标注来获得图像高层语义概念的标注,实现待标注图像语义的自动标注。我们将图像的语义可以定义为属性概念和高层抽象概念,采用二次标注方法实现对于图像语义的自动标注。实验证明,本文的方法可以使图像获得丰富的高层抽象语义概念标注,从而缩小"语义鸿沟",有效提高了检索的效率和精确度。更多还原

【Abstract】 A new image annotation framework using ontology is presented in this paper.We utilized the hierarchical probabilistic image concept reasoning for high-level image annotation,which combined with the relationship between concepts in domain ontology,to achieve automatic image semantic annotation.We can define the semantics of the image concept as image attribute concept and highlevel abstract concept,then use two steps to achieve the automatic semantic annotation for the image.The experiment results show that it makes the image obtain the excellent high-level abstract semantic concept annotations,thus reducing the "semantic gap ",and effectively improving the efficiency and accuracy of retrieval.更多还原