基于DCT域特征和Bayes语义分类的教学图像资源库检索技术  

Research on Retrieving Image Resource Technique Based on Bayes Semantic Classification and Visual Feature Extraction

作　　者：侯洁[1] 

机构地区：[1]天津外国语大学教育技术与信息学院,天津300204

出　　处：《新疆师范大学学报（自然科学版）》2012年第4期54-61,共8页Journal of Xinjiang Normal University(Natural Sciences Edition)

摘　　要：提出了一种根据DCT域特征提取和Bayes语义分类结合的图像检索技术。针对图像库数据结构的语义需求,采用朴素Bayes算法对图像库进行分类;其中视觉特征提取部分,采用改进的DCT域的边缘空间分布概率特征提取方法,获取目标对象边缘信息,建立20个特征值;通过Bayes分类器的分类,获取每类的索引特征向量,用于初次语义类别过滤,提高了检索效率。通过实验,建立教学需要的图像资源库,与其它检索算法进行查准率对比试验,证明了本算法的优越性。A image retrieval method was proposed based on feature extracted in DCT and semantic classification by bayes. Firstly, Semantic need for retrieving digital image was studied, and Simple Bayesian Classifier was used in semantic classification of image resource; Secondly, improved image feature of edge space distribution probability based on DCT is extracted to obtain edge information of goal object and set up 20 Eigen values. Thirdly, indexing feature vectors are obtained through semantic classification for semantic filtration, by which the retrieval efficiency is improved. With experiment, image resource data- base is established and the proposed algorithm shown better result by experiment on precision comparison compared with other methods.

关 键 词：图像检索 BAYES分类 DCT域特征提取 图像语义分类 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]