改进的BEMD在纹理图像检索中的应用

二维经验模式分解（BEMD）方法把图像分解成一组称为固有模式函数（IMF）和残差,广泛应用在纹理图像的检索中。对BEMD检索纹理图像进行改进,使用快速自适应二维经验模式分解（FABEMD）对纹理图像进行分解,创新性地利用广义伽马分布密度函数来对FABEMD分解出来的IMF系数进行建模,并利用Kullback-Leibler距离来计算两张纹理图像的相似性。实验结果表明,算法快速有效。

Application of the Improved BEMD Method in Texture Image Retrieval

The BEMD method decompose image into a set of functions named IMF and residue.It is widely used in the texture image retrieval. Improves the BEMD method, uses the FABEMD to decompose images. The Generalized Gamma Density is used to model the coefficients of IMFs and the Kullback-Leibler Distance is used to compute the two texture image similar. Experiments show that the improved BEMD is more faster and more effective.