提出了一种complex moment的圆度测量方法.作为图像的形状特征,该方法易于计算,并可以减少形状提取时的大量计算,对噪声和狭窄区域有很好的鲁棒性.使用图像结构相似度度量方法FSIM,结合complexmoment的圆度特征、图像的相位特征、相位不变性特征,既弥补相位不变性(PC)对图像对比度不变的影响,又能充分利用圆度测量计算简单,旋转、放缩不变性,计算量小的优点.为了验证该算法效率,采用图像库Shape7,彩色物体库COL、VIEW、ILL进行图像检索实验.实验表明,该方法对图像的形状和明度变化有很好的检索效果.%This paper proposed a new algorithm based on circularity measure of complex moment as a shape measure. An addi-tional advantage of the generalized measure was that it could be used for reducing the compulation cost and detecting small ir-regularities in nearly circular shapes damaged by noise or during an extraction process. And it used FS1M as measure distance with circularity, phase congruent and gradient magnitude. This method could not only fill up a deficiency of the contrast invariant of phase congruent, but also fully utilize the scaling and rotation invariant of circularity. To evaluate the effectiveness of the proposed method.it carried out a series of experiments on one Shape7 database and three color image databases COL, VIEW and ILL. The experiments show that the proposed method has higher retrieval precision rate and is more efficient for the different contrast and shape image.
展开▼
