基于梯度和局部多值模式的图像纹理特征提取  

Image Texture Feature Extraction Based on Gradient and Local Multi-Valued Pattern

作　　者：赵珊[1] 于虎[1] 

机构地区：[1]河南理工大学计算机科学与技术学院,河南焦作454000

出　　处：《测控技术》2017年第8期6-10,共5页Measurement & Control Technology

基　　金：国家自然科学基金项目(51274088);河南省基础与前沿技术研究项目(132300410462)

摘　　要：为了进一步增强局部二值模式的纹理特征表达能力,提出了一种将图像梯度同局部多值模式相结合的纹理特征提取方法。该方法首先计算出图像的梯度幅度和梯度方向,并根据量化的梯度方向对幅度图进行分解,之后利用得到的幅度图对其进行局部多值模式特征的提取,得到一种新的纹理特征,最后将得到的特征进行图像检索实验。实验结果表明,在纹理丰富的检索库上,该方法所提取的纹理特征更加具有通用性,相对传统局部二值模式图像检索方法具有更好的稳健性和更好的检索效果。In order to enhance the texture feature expression of the local binary pattern, an image texture fea- ture extraction method combining image gradient and local multi-valued pattern is proposed. Firstly, the gradi- ent amplitude and gradient direction of the image are calculated, and the magnitude map according to the quan- tized gradient direction is decomposed, then the local features with the multi-valued pattern on the amplitude image are extracted, and a new texture feature is obtained. Finally, an image retrieval experiment is carried out with this feature. The experimental results show that the texture feature extracted by this method is more univer- sal, and has better robustness and retrieval effect than traditional local binarv pattern in rich texture datahase.

关 键 词：局部二值模式 纹理特征 图像梯度 图像检索 特征表达 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]