基于二进小波直方图的纹理图像检索

为了进一步提高小波直方图的检索性能，提出了一种二进小波直方图算法。该算法根据二进小波的特点，从能量角度出发．通过子带组合，0／1量化等运算生成二进小波直方图。与单小波直方图相比，二进小波利用模局部极大值来提取图像的多尺度边缘信息并具有时移不变的特性。因此二进小波直方图具有特征提取快和检索精度高等特点．纹理图像检索的对比实验结果表明二进小波直方图的检索精度比单小波直方图提高了12％。

Dyadic Wavelet Histogram for Texture Image Retrieval

In order to enhance the performance of the scalar wavelet histogram, a dyadic wavelet histogram algorithm was presented, which had obtained by combining the dyadic wavelet channels and 0/1 quantity from the energy of wavelet coefficients using the characteristics of the dyadic wavelet decomposition. In comparison with the scalar wavelet histogram, the dyadic wavelet have local modulus maximum of a wavelet transform to find multiscale edges information and has shift invariant and redundancy, so the method is better than the scalar wavelet histogram, and has some advantages such as less cost of computing, high retrieval rate. In the contrasted experiment result for the texture image retrieval, the retrieval efficiency is raised 12%.