Deep Learning-based systems for Finger vein recognition have gained rising attention in recent years due to improved efficiency and enhanced security.The performance of existing CNN-based methods is limited by the pun...Deep Learning-based systems for Finger vein recognition have gained rising attention in recent years due to improved efficiency and enhanced security.The performance of existing CNN-based methods is limited by the puny generalization of learned features and deficiency of the finger vein image training data.Considering the concerns of existing methods,in this work,a simplified deep transfer learning-based framework for finger-vein recognition is developed using an EfficientNet model of deep learning with a self-attention mechanism.Data augmentation using various geometrical methods is employed to address the problem of training data shortage required for a deep learning model.The proposed model is tested using K-fold cross-validation on three publicly available datasets:HKPU,FVUSM,and SDUMLA.Also,the developed network is compared with other modern deep nets to check its effectiveness.In addition,a comparison of the proposed method with other existing Finger vein recognition(FVR)methods is also done.The experimental results exhibited superior recognition accuracy of the proposed method compared to other existing methods.In addition,the developed method proves to be more effective and less sophisticated at extracting robust features.The proposed EffAttenNet achieves an accuracy of 98.14%on HKPU,99.03%on FVUSM,and 99.50%on SDUMLA databases.展开更多
In this paper, we propose a novel method for finger-vein recognition. We extract the features of the vein patterns for recognition. Then, the minutiae features included bifurcation points and ending points are extract...In this paper, we propose a novel method for finger-vein recognition. We extract the features of the vein patterns for recognition. Then, the minutiae features included bifurcation points and ending points are extracted from these vein patterns. These feature points are used as a geometric representation of the vein patterns shape. Finally, the modified Hausdorff distance algorithm is provided to evaluate the identifica-tion ability among all possible relative positions of the vein patterns shape. This algorithm has been widely used for comparing point sets or edge maps since it does not require point cor-respondence. Experimental results show these minutiae feature points can be used to perform personal verification tasks as a geometric rep-resentation of the vein patterns shape. Fur-thermore, in this developed method. we can achieve robust image matching under different lighting conditions.展开更多
针对近红外光下采集的指静脉图像存在局部像素相似性强、单一方向识别效果差的问题,提出模板投票和多方向融合的指静脉识别方法。首先,提出基于模板投票的局部三值模式(template voting local three pattern,TVTP),充分利用局部多邻域...针对近红外光下采集的指静脉图像存在局部像素相似性强、单一方向识别效果差的问题,提出模板投票和多方向融合的指静脉识别方法。首先,提出基于模板投票的局部三值模式(template voting local three pattern,TVTP),充分利用局部多邻域点的信息,减少局部像素相似性;其次,指静脉图像中含有丰富的方向特征信息,提出多方向编码(multi-directional coding,MDC),获取图像中具有辨别力的方向特征,加强不同方向特征之间的鲁棒性,解决单一方向识别率差的问题;最后,利用分块直方图统计特征,并使用协同表示(collaborative representation,CR)进行分类,提高识别效率。实验结果证明,所提方法在SDUMLA数据集、USM数据集和THU-FVFDT2数据集上的识别率分别达到99.32%、99.73%和99.75%,与其他经典和新颖算法相比,不仅取得了更好的识别效果,还能同时满足实时性要求,具有应用价值。展开更多
文摘Deep Learning-based systems for Finger vein recognition have gained rising attention in recent years due to improved efficiency and enhanced security.The performance of existing CNN-based methods is limited by the puny generalization of learned features and deficiency of the finger vein image training data.Considering the concerns of existing methods,in this work,a simplified deep transfer learning-based framework for finger-vein recognition is developed using an EfficientNet model of deep learning with a self-attention mechanism.Data augmentation using various geometrical methods is employed to address the problem of training data shortage required for a deep learning model.The proposed model is tested using K-fold cross-validation on three publicly available datasets:HKPU,FVUSM,and SDUMLA.Also,the developed network is compared with other modern deep nets to check its effectiveness.In addition,a comparison of the proposed method with other existing Finger vein recognition(FVR)methods is also done.The experimental results exhibited superior recognition accuracy of the proposed method compared to other existing methods.In addition,the developed method proves to be more effective and less sophisticated at extracting robust features.The proposed EffAttenNet achieves an accuracy of 98.14%on HKPU,99.03%on FVUSM,and 99.50%on SDUMLA databases.
文摘In this paper, we propose a novel method for finger-vein recognition. We extract the features of the vein patterns for recognition. Then, the minutiae features included bifurcation points and ending points are extracted from these vein patterns. These feature points are used as a geometric representation of the vein patterns shape. Finally, the modified Hausdorff distance algorithm is provided to evaluate the identifica-tion ability among all possible relative positions of the vein patterns shape. This algorithm has been widely used for comparing point sets or edge maps since it does not require point cor-respondence. Experimental results show these minutiae feature points can be used to perform personal verification tasks as a geometric rep-resentation of the vein patterns shape. Fur-thermore, in this developed method. we can achieve robust image matching under different lighting conditions.
文摘针对近红外光下采集的指静脉图像存在局部像素相似性强、单一方向识别效果差的问题,提出模板投票和多方向融合的指静脉识别方法。首先,提出基于模板投票的局部三值模式(template voting local three pattern,TVTP),充分利用局部多邻域点的信息,减少局部像素相似性;其次,指静脉图像中含有丰富的方向特征信息,提出多方向编码(multi-directional coding,MDC),获取图像中具有辨别力的方向特征,加强不同方向特征之间的鲁棒性,解决单一方向识别率差的问题;最后,利用分块直方图统计特征,并使用协同表示(collaborative representation,CR)进行分类,提高识别效率。实验结果证明,所提方法在SDUMLA数据集、USM数据集和THU-FVFDT2数据集上的识别率分别达到99.32%、99.73%和99.75%,与其他经典和新颖算法相比,不仅取得了更好的识别效果,还能同时满足实时性要求,具有应用价值。
