基于局部线性二值模式的掌纹掌脉融合识别

为了改善图像模糊导致鲁棒性下降这一问题,提出了一种基于掌纹掌脉双模态融合的多生物特征识别方法.对传统的线性二值模式(LBP)方法进行改进,提出了一种新的特征提取方法——局部线性二值模式(LLBP);对掌纹掌脉图像进行分块操作,形成分块局部线性二值模式(BLLBP),并将掌纹和掌脉特征进行融合;利用汉明距离进行匹配.实验图库分别为接触式公用图库和自建PolyU模糊图库以及自建SUT-D模糊图库,并与目前典型方法进行对比实验,结果表明:在三个图库中可分别获得最低等错误率(EER)为1.1513%,4.5162%和7.0439%,充分证明了该方法能够进一步提高身份识别的防伪性、鲁棒性及识别精度,具有可行性.

Palmprint and palm vein fusion recognition based on LLBP

In order to improve the problem that blur of images resulting in worse robustness,a multi-biometric characteristic identification method is proposed based on dual mode fusion of palmprint and palm vein.Traditional linear binary pattern(LBP)is improved and a new feature extraction method called local LBP(LLBP)is proposed, the palmprint and palm vein image is divided into blocks,to form block LLBP(BLLBP),feature fusion is implemented,matching is carried out by Hamming distance.The experiment databases are contact public database, self-built PolyU blur map storage and self-built SUT-D blur map storage respectively.The comparison experiments were implemented with the current typical approaches in this three databases.The result of experimental show that the method can obtain the lowest EER,which is 1.151 3%,4.516 2% and 7.043 9% respectively.It shows that the method can further improve the security,robustness and recognition accuracy of identity recognition,which is feasible.