一种三结构描述子的图像检索方法
发布时间：2018-11-27 14:26
【摘要】：针对基元结构描述子在颜色空间的基础上再提取其他特征,导致偏重对颜色信息的描述而降低了图像检索性能的问题,提出一种应用在HSV颜色空间上的三结构描述子(TSD)的特征提取方法。该方法在HSV颜色空间中分别提取颜色和纹理信息,考虑到了颜色和纹理特征的同等重要性,同时避免了颜色信息的过多干扰;在纹理特征提取中,TSD利用像素间的信息变化来表示局部空间结构信息,解决了传统的局部模式方法忽略对局部结构的空间关系描述的问题,获得了更多的空间结构信息。实验结果表明,该方法在3个图像库Corel-1000、Corel-5000和Corel-10000上的检索准确率分别达到78.08%、38.12%和52.12%,与以往基元结构方法相比,检索准确率得到了提高。
[Abstract]:The primitive structure descriptor extracts other features on the basis of color space, which leads to the problem that the description of color information is biased and the performance of image retrieval is reduced. This paper presents a feature extraction method for (TSD), a three-structure descriptor applied to HSV color space. The method extracts color and texture information in HSV color space, considering the equal importance of color and texture features, and avoids the excessive interference of color information. In texture feature extraction, TSD uses the information change between pixels to represent the local spatial structure information, which solves the problem that the traditional local pattern method ignores the spatial relationship description of the local structure and obtains more spatial structure information. The experimental results show that the retrieval accuracy of this method on Corel-1000,Corel-5000 and Corel-10000 is 78.08% and 52.12%, respectively. Compared with the previous primitive structure method, the retrieval accuracy is improved.
【作者单位】： 大连理工大学计算机科学与技术学院;大连理工大学创新创业学院;大连理工大学控制科学与工程学院;
【基金】：国家自然科学基金资助项目(61370200,61672130,61602082)
【分类号】：TP391.41
本文编号：2361112
[Abstract]:The primitive structure descriptor extracts other features on the basis of color space, which leads to the problem that the description of color information is biased and the performance of image retrieval is reduced. This paper presents a feature extraction method for (TSD), a three-structure descriptor applied to HSV color space. The method extracts color and texture information in HSV color space, considering the equal importance of color and texture features, and avoids the excessive interference of color information. In texture feature extraction, TSD uses the information change between pixels to represent the local spatial structure information, which solves the problem that the traditional local pattern method ignores the spatial relationship description of the local structure and obtains more spatial structure information. The experimental results show that the retrieval accuracy of this method on Corel-1000,Corel-5000 and Corel-10000 is 78.08% and 52.12%, respectively. Compared with the previous primitive structure method, the retrieval accuracy is improved.
【作者单位】： 大连理工大学计算机科学与技术学院;大连理工大学创新创业学院;大连理工大学控制科学与工程学院;
【基金】：国家自然科学基金资助项目(61370200,61672130,61602082)
【分类号】：TP391.41
【相似文献】
相关硕士学位论文 前1条
1 刘志忠;基于谱图理论的结构描述子及其在点模式匹配中的应用[D];安徽大学;2014年
，本文编号：2361112
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2361112.html
