【作者】 张玲；

【导师】 孙家广； 邹北骥；

【作者基本信息】 湖南大学 ， 通信与信息系统， 2006， 硕士

【摘要】 基于内容的图像检索是根据描述图像视觉内容的特征向量进行相似性检索,其中图像视觉内容的提取可以是通用的,也可以是基于特定领域的。基于特定领域的图像检索技术可以充分利用该领域内的有用知识,其检索结果更能符合用户的需求。随着市场经济的发展,注册商标的数目不断增加,需要建立一种准确、高效的商标图像自动检索系统,因此基于内容的图像检索技术在商标检索领域得到了非常广泛的应用。本文对基于内容的商标图像检索,尤其是对分块检索方法进行了研究,提出了一种新的基于极坐标下分块的商标图像分块检索方法,并设计开发了一个基于综合特征检索的商标图像检索系统。商标图像的分块检索是将商标图像分块后,利用全局和局部图像特征来检索商标图像,每个子块图像特征能够反映图像的局部特征,而多个子块图像特征的结合又能够对区域的整体形状进行描述。针对现有的四叉树分块方法的不足,提出了一种极坐标下的分块检索方法。首先对商标图像进行预处理,得到规则二值图像;计算目标像素的最小外接圆和图像的形状主方向;在以最小外接圆圆心为原点、形状主方向为极轴方向的极坐标中将图像分块,提取各子块特征,构成形状直方图并平滑处理;再结合全局形状特征来综合表示图像内容;最后采用新的相似性度量方法计算图像间距离,实现商标图像的准确检索。实验证明,该方法具有良好的平移、缩放、旋转和镜像不变性,得到的检索结果能很好地符合人的视觉感受。本文设计开发了一套基于综合特征检索的商标图像检索系统,其目标是一个实验性的框架系统,主要是作为各种检索算法的测试平台,可以很方便地定制不同的分块检索算法。更多还原

【Abstract】 Content based image retrieval is to perform the similarity retrieval according to the image features representing the image content, which may be extracted in the generic or specific domain. Specific domain based image retrieval can benefit from the useful domain knowledge and the corresponding retrieval results can meet the user’s requirement better than the generic one. With the development of market economy the trademark plays the more and more important role in the current society and the content based image retrieval technique has been widely used in the trademark domain. This paper does research mainly on content based trademark image retrieval, especially on subblock-based image retrieval, presenting a new subblock-based image retrieval method under the polar coordinate system for retrieving trademark image, and designs trademark tmage retrieval system based on integrated features.The subblock-based image retrieval uses global and local image features to retrieve trademark image. The image is partitioned into several subblocks, Each subblock image features reflect the image local information and the combination of multiple subblocks’ image features describe the whole region shape. This paper proposes a subblock-based image retrieval method under the polar coordinate system for retrieving trademark image. After preprocessing the trademark image, the regular binary-image is obtained. Then, the minimum circumscribed circle and shape principal orientation of the object region are calculated; and the object region is divided into some blocks in the polar coordinate, which uses the center of the minimum circumscribed circle as the origin and the direction of the shape principal orientation as the direction of the polar axis,and the features of all subblocks are extracted to compose shape histograms and then smoothed; Then, combined with global features to synthetically represent the content of the image; Finally, a new image similarity is used to compute the distance between images. Experiment results show that this method can keep good invariance under translation, rotation, scale and mirror, and the retrieved results match human visual perception very well.This paper designs a trademark image retrieval system based on integrated features, which is aimed to be an experimental system and can be used to test different retrieval algorithms. It is also very convenient to customize the different subblock-based retrieval methods.更多还原