.png){kind=icon}
返回
Learning Discriminant Direction Binary Palmprint Descriptor
.png){kind=icon}
PRE
AI
DOI：10.1109/TIP.2019.2903307.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Palmprint directions have been proved to be one of the most effective features for palmprint recognition. However, most existing direction-based palmprint descriptors are hand-craft designed and require strong prior knowledge. In this paper, we propose a discriminant direction binary code (DDBC) learning method for palmprint recognition. Specifically, for each palmprint image, we first calculate the convolutions of the direction-based templates and palmprint and form the informative convolution difference vectors by computing the convolution difference between the neighboring directions. Then, we propose a simple yet effective model to learn feature mapping functions that can project these convolution difference vectors into DDBCs. For all training samples: (1) the variance of the learned binary codes is maximized; (2) the intra-class distance of the binary codes is minimized; and (3) the inter-class distance of the binary codes is maximized. Finally, we cluster the block-wise histograms of DDBC forming the discriminant direction binary palmprint descriptor for palmprint recognition. The experimental results on four challenging contactless palmprint databases clearly demonstrate the effectiveness of the proposed method.
Keyword:
Palmprint recognition
direction feature learning
discriminant direction binary code
biometrics
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
13.7
论文数:
1.0W
被引数:
8.4W
机构
引用论文
分享收藏
分享收藏Multi-spectral palmprint recognition based on oriented multiscale log-Gabor filters.png)
NEUROCOMPUTING
OAAI
NEUROCOMPUTINGIF6.5
OAAI分享收藏
查看更多内容