改进的LTP与2DPCA结合的掌纹识别

Combination of Improved LTP and 2DPCA for Palmprint Recognition

本文以将掌纹作为研究的对象,提出了改进的局部三值模式(Local Ternary Pattern,LTP)与二维主分量分析(Modular 2 Dimensional Principle Component Analysis,2DPCA)相结合的掌纹识别方法。该方法先对原始的掌纹图像利用改进的局部三值模式方法提取特征,然后利用二维主分量分析的方法对提取的特征进行降维,同时消除特征之间的冗余性,最后用欧氏距离判别法进行掌纹的识别。在Poly U标准数据库中进行实验,结果显示,在相同的训练样本下,改进的LTP与2DPCA结合的方法相比于LTP方法,大大降低了特征维数且具有更高的正确识别率。从而证明了该方法的有效性。

This paper will treat palm print as the object of study, presenting a palm print identification approach, which was the combination of improved LTP and 2DPCA operation. Improved LTP is first used to get the pixel values of palm print diagram and used as the original features of palm print image, and then 2DPCA method is used to eliminate the high redundancy in the characteristics extracted, which effectively reduced the dimensions of the characteristics, and then Euclidean distance was used to implement the palm print image classification. Finally, experiment results based on the standard library PolyU. It experimented on the same training images, the results show that compared to LTP method, the method of combination of improved LTP and 2DPCA shows its advantage. It reduces feature dimensions, at the same time, it also improves the correct recognition rate. This method demonstrates its effectiveness.