利用无参数统计特征进行旋转不变纹理图像渐进检索

A Rotation-invariant Texture Retrieval Algorithm Based on Parameterless Statistical Features

分析了Radon变换和Log-polar变换在消除旋转位移时对频谱的影响,探讨了NSCT变换和小波变换在不同检索参数下的平均检索性能,在此基础上构造出多尺度多方向纹理变换谱和旋转不变特征矢量,并提出了一种基于无参数统计特征的旋转不变纹理图像两级渐进检索算法。通过VisTex标准纹理图像库的检索实验证明,与传统检索算法相比,本文提出的算法既可获取纹理主方向,同时又能表征纹理细节信息,有效地提高了旋转不变纹理图像检索的查准率和检索效率。

We analyze the spectrum influence between Radon transform and Log-polar transform when rotation effect is eliminated.The average retrieval performance of wavelet and NSCT with different retrieval parameters is also studied.Based on which,we design a multi-scale and multi-orientation texture transform spectrum,as well as rotation-invariant feature vector and its measurement criteria.Then a new two-level rotation-invariant texture retrieval algorithm based on no-parameter statistic features is proposed.The experimental results on Brodatz image database show that our algorithm is appropriate for main orientation capturing and detail information description.The combination of this two-level progressive retrieval strategy and multi-scale analysis method can effectively improve retrieval efficiency,compared with traditional algorithms,and ensure a high precision as well.