基于小波分解和Hu不变矩的图像检索算法

Image Retrieval Algorithm Based on Wavelet Decomposition and Hu Invariant Moment

提出了一种基于小波分解和Hu不变矩的图像检索算法。首先,对每幅图像进行三级小波分解,得到图像的低频信息和高频信息,把分解得到的低频信息部分的Hu不变矩和各个高频信息部分的均值与方差作为图像的特征向量,利用欧氏距离进行相似度度量,实现基于内容的图像检索。为了验证文中所提到算法的检索效果,用查准率作为评价标准,分别与基于Contourlet变换的图像检索算法和基于Hu不变矩的图像检索算法进行了比较。实验结果表明,图像检索算法效率高,不仅能使检索结果的排序值得到改善,而且还有较高的查准率。

This paper presents an image retrieval algorithm based on wavelet decomposition and Hu invariant moments.Firstly,the three-level wavelet decomposition is carried out for each image,and the low-frequency information and high-frequency information of the image are obtained.The Hu-invariant moments of the low-frequency information parts and the mean and variance of each high-frequency information part are taken as the eigenvectors of the image.Euclidean distance is used to measure similarity to realize content-based image retrieval.In order to verify the retrieval effect of the proposed algorithm,the accuracy of the algorithm is used as the evaluation criterion,and compared respectively with the image retrieval algorithm based on Contourlet transform and the image retrieval algorithm based on Hu invariant moments.The experimental results show that the proposed image retrieval algorithm is efficient and can not only improve the sorting value of the retrieval results,but also have a high precision.