摘要
提出了一种新型快速旋转不变图像检索新方法。该方法首先对图像进行傅里叶变换和功率谱分解,提取功率谱的扇形区域能量和环形区域能量参数,并将其均值和标准差作为图像纹理特征。然后,利用谱能量分布特征把纹理的主方向旋转到0,°提取旋转后图像的共生矩阵参数和小波分解各子带图像统计参数作为基本特征。利用所提出的特征提取方法在两组分别包含25类单色自然纹理的图像库上进行检索试验。结果表明,该方法获得了良好的检索效果。
A novel fast rotation invariant image retrieval method was proposed. First, Fourier transform and Fourier spectrum decomposition were performed. Sector energy and ring energy of spectrum were extracted and their mean and standard deviation were calculated as texture features. Then, texture image was rotated to place its dominant direction at 0 degree according to the spectral energy distribution. Co-occurrence matrix based features and wavelet statistical features of the rotated image were calculated as basic texture features. Image retrieval using the proposed method showed a high performance conducted with two different texture databases both of which include 25 kinds of monochromatic natural textures.
出处
《计算机应用研究》
CSCD
北大核心
2007年第4期166-169,共4页
Application Research of Computers
基金
国家"863"发展计划资助项目(2005AA121520)
国家自然科学基金资助项目(60272014)
关键词
图像检索
谱分解
小波变换
image retrieval
spectrum decomposition
wavelet transform
