基于改进双树复小波和灰度-梯度共生矩阵的纹理图像检索算法

Retrieval Algorithm for Texture Image Based on Improved Dual Tree Complex Wavelet Transform and Gray Gradient Co-occurrence Matrix

针对双树复小波变换缺少不同尺度纹理的空间分布特征的缺陷,提出了一种改进双树复小波和灰度-梯度共生矩阵相融合的纹理图像检索新算法。首先,该算法将图像进行非均匀分块,并对分块的图像进行双树复小波变换,以此增加不同尺度下的空间信息;其次,利用灰度-梯度共生矩阵提取4个统计量特征;然后,融合两种方法提取的纹理特征以得到图像检索的纹理特征;最后,用Canberra距离进行相似性度量并输出图像检索的结果。实验结果表明,该方法对纹理图像有较好的检索效果。

To overcome the lack of texture space distribution characteristics at different scales for dual tree complex wavelet transform,a kind of improved retrieval algorithm for texture image was proposed based on the combining dual tree complex wavelet transform with gray gradient co-occurrence matrix.Firstly,in order to increase the spatial information at different scales,the algorithm will be non-uniform image blocks,and every block image is dual tree complex wavelet transform.Secondly,the four statistical features are extracted by using gray gradient co-occurrence matrix.Then,texture features for image retrieval are obtained using fusion of texture feature extracted by above two methods.Finally,Canberra distance is used similarity measure and output image search results.Experimental results show that this method has better retrieval results for texture images.