摘要
提出一种新的利用对象本体进行基于内容的图像检索的方法。从分割后的图片区域中提取低层特征,并将其映射为本体中的中间层描述符。中间层描述符将图像低层特征与高层语义联系起来,实现基于内容的图像检索。实验证明,该方法适于大型图片库的检索。
This paper presents an novel approach to content-based image retrieval using ontology technology. Low-level features were extracted from regions after image segmentation and then mapped to middle-level descriptors which form the ontology. Middle-level descriptors enable the transformation from low-level features to high-level semantic and made CIBR easier. Experiments show the method was promising.
出处
《计算机与数字工程》
2007年第7期111-113,147,共4页
Computer & Digital Engineering
基金
国家社会科学基金
(编号02CTQ002)资助。
