| 摘要: |
| 针对伪造的手指静脉图像能够成功攻击手指静脉识别系统，从而使得其识别系统安全性能大大降低的问题，提出了一种基于深度置信网络的手指静脉防伪检测的方法；通过逐层无监督的学习方法预训练深度网络的权值参数，以及有监督的BP神经网络微调深度网络的权值参数，从而提取到手指静脉图像的特征，用于静脉图像的检测；实验结果证明所提出的手指静脉防伪检测方法能够有效地识别出假手指静脉图像；通过对比性实验研究，发现此方法提高了手指静脉识别系统的安全性能。 |
| 关键词: 手指静脉 防伪检测 深度学习 深度置信网络 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Fake Finger Vein Image Detection Based on Deep Belief Network |
|
LIU Xiaa， QIN Hua fengb
|
| Abstract: |
| According to that fake finger vein images can successfully attack finger vein recognition system， making the safety performance significantly reduced， this paper proposes a novel method for finger vein spoofing attack detection based on Deep Belief Network. Through the unsupervised training layer by layer to get the weight parameters of the depth network and the supervised BP neural network to fine tune the weight parameters， we extract the feature of the finger vein image and detect the vein image. Obtained experiment results show that the proposed scheme can effectively identify the fake finger vein image. Through comparative experimental research， we find that the proposed scheme can improve the safety performance of the recognition system. |
| Key words: finger vein anti spoof detection deep learning deep belief network |
