基于纹理语义特征的图像检索研究

Research on Texture()-Based Semantic Image Retrieval

针对Tamura纹理模型,提出了基于语言变量的图像纹理语义特征描述方法;并通过遗传程序设计构造从低层视觉特征到高层语义特征的映射;最后根据这些模糊语义值进行图像检索.实验结果表明系统不仅能得到出众的检索效率,而且与人类的视知觉具有比较好的一致性,提出的方法对于缩小低层视觉特征和高层语义特征之间的“语义鸿沟”具有很大的意义.

According to Tamura＇s texture model, this paper puts forward an image texture semantic description framework based on linguistic variable. Authors successfully construct the mapping from low-level visual feature to high-level semantic feature through genetic programming algorithm, and propose the fuzzy retrieval algorithm according to the extracted semantic features. The experiment results show that above approach not only has an excellent retrieval precision but also has a good accordance with the human visual perception. The approach has strong significance for reducing the ＂semantic gap＂ between the visual feature and semantic visual.