中文摘要：

为提高图像检索效率，本文提出一种Fourier-Mellin矩的彩色图像检索方法．该方法利用彩色图像的Hilbert变换下相位一致性和梯度幅值的结构特征，用Fourier-Mellin矩描述图像纹理特征．既利用相位一致性不受图像亮度和对比度变化的特点，又利用Fourie-Mellin矩的旋转不变性的特点，对彩色图像有很好的检索效果．为验证该算法的效率，在三个彩色图像库中进行对比实验，实验表明：该算法明显优于文中比较算法．

英文摘要：

In order to improve the effect of the color image retrieval, a new algorithm is proposed based on Fourier- Mellin moments, phase congruent and gradient magnitude in this paper. This method aims to extract phase congruent, gradient magnitude of an image smoothed with Hilbert transform, and Fourie~Mellin moments, which not only can fill up an efficiency of the contrast invariant of phase congruent, but also can fully utilize the rotation invariant of Fourier-Mellin moments. The retrieval experiments are conducted in three color databases to test the performance of our method. The experiments show that our method in this paper is more efficient comparing to traditional methods.