一种新的空间颜色直方图及其度量方法

A Novel Spatial Color Histograms and Similarity Measurement Method

针对传统颜色直方图存在维数高、受光照影响、相近颜色缺少相关性以及丢失空间信息的问题,提出一种基于聚类的空间颜色直方图方法.该方法首先对图像进行k-均值聚类,然后在聚类图上统计基于空间位置分布的颜色直方图.利用pairwise参数学习方法学习出HSV颜色空间的距离度量参数,并提出基于最近颜色对迭代的空间颜色直方图的相似性度量算法.将基于聚类的空间颜色直方图特征用于图像检索实验,并与基于其它颜色直方图检索方法比较,查准率和查全率均有所提高.结果表明,该方法较好地描述了图像的主、客观颜色特征,具有较强的适应性和鲁棒性.

There are four problems on color histograms： the dimension of color histograms is too large; color histograms are easy affected by the change of illumination; similar colors are not relevant; and color histograms lack spatial information. A novel algorithm of cluster-based spatial color histograms is proposed to solve these problems. Firstly, this method runs k-means cluster on the source image, and constructs spatial color histograms on the image created by k-means cluster. We train the parameters of HSV-space distance formula by using a pariwise-based learning method, and then propose a histogram matching algorithm based on the most similar colors iteration. Our cluster-based spatial color histograms feature is used in image retrieval experiments, compared with other color histograms based image retrieval methods, precision and recall are improved. The results show that our method can describe subjective and objective image color features very well, and has stronger adaptation and robustness.