中文摘要：

基于内容的图像检索在很多领域都有广泛的应用．传统方法通常提取图片的底层视觉特征，如颜色、纹理和形状等，进行在视觉特征空间下的相似查询．然而，这些视觉特征无法表达图片需要传递的情感和概念信息．提出一种基于视觉和主观特征的统一图像概率检索方法．具体来说，图片通过三种类型特征（视觉特征、风格特征和情感特征等）来表达．通过线性加权得到图片间的统一相似距离，其中权重参数通过多元回归得到．不同于常规图像检索方法，只采用视觉相似度作为查询的相似尺度，该方法允许用户选择三种特征作为查询元素，而且，通过引入概率模型实现对检索结果进行一定置信度保证的进一步细化．实验表明该检索方法及其索引的有效性．

英文摘要：

Content-based image retrieval （ CBIR ） has been extensively studied due to its wide application in many fields. In the state- ofthe-art methods, color, texture and shape are extracted from an image as low level visual features and then search is conducted over the visual feature space. These visual features, however, can not express the sentiment concepts that an image can deliver. To address this issue, in this paper, we present a effective image retrieval by conducting the search over combined visual and conceptual feature spaces. Specifically, three types features, visual, style and sentiment, are introduced to represent an image. An unified similarity distance between two images is obtained by linearly concatenating the three similarity measures over three feature spaces, where weight parameters are obtained by a multi-variable regression method. Different from conventional image retrieval methods which only adopt visual similarity as a query metric, our proposed retrieval algorithm allows user to choose the above three kinds of features that they prefer to as query elements. Moreover, a probabilistic modal is introduced to refine the retrieval result with confidence guarantee. Comprehensive experiments are conducted to testify the effectiveness and efficiency of our proposed retrieval and indexing methods respectively.