摘要
针对图像的底层特征与高层语义特征之间建立映射,使用基于支持向量机(SVM)的语义关联方法,将HSV颜色特征作为SVM的输入参数,对图像库学习和分类,建立图像底层特征与高层语义的关联,并结合图像底层特征和语义信息进行检索。实验表明:该方法提高了检索效率,取得了较高的准确率。
This paper presents a method that builds a map between image＇s low level visual features and high level semantic features based on SVM.We use the HSV color features as input data,learn and classify the images database,build relation and mapping between image＇s low level features and the semantics,and combine the low level features and the semantic information.Laboratory tests show that efficiency and accuracy are improved.
出处
《上海电机学院学报》
2010年第6期343-346,352,共5页
Journal of Shanghai Dianji University
基金
上海电机学院重点学科资助(07xkj01)
上海电机学院科研启动经费项目(10C415)
关键词
支持向量机
语义检索
视觉特征
support vector machine（SVM）
semantic retrieval
visual feature
