[1]张 浩,陈亚南,杨 璐,等.基于注意力机制与残差网络的掌纹脊线距离估计[J].计算机技术与发展,2023,33(07):61-67.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 009]
　ZHANG Hao,CHEN Ya-nan,YANG Lu,et al.Palmprint Ridge Distance Estimation Based on Attention Mechanism and Residual Network[J].,2023,33(07):61-67.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 009]

点击复制

基于注意力机制与残差网络的掌纹脊线距离估计()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年07期

页码:

61-67

栏目:

媒体计算

出版日期:

2023-07-10

文章信息/Info

Title:

Palmprint Ridge Distance Estimation Based on Attention Mechanism and Residual Network

文章编号:

1673-629X(2023)07-0061-07

作者:

张 浩; 陈亚南; 杨 璐; 许丽娜; 郝凡昌

山东建筑大学 计算机科学与技术学院,山东 济南 250101

Author(s):

ZHANG Hao; CHEN Ya-nan; YANG Lu; XU Li-na; HAO Fan-chang

School of Computer Science and Technology,Shandong Jianzhu University,Jinan 250101,China

关键词:

掌纹脊线距离; VGG16; 残差网络; 注意力机制; 损失函数

Keywords:

palmprint ridge distance; VGG16; residual network; attention mechanism; loss function

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 07. 009

摘要:

掌纹脊线距离是掌纹脊线的重要纹理属性之一,在掌纹识别算法与图像质量评估方法中常作为一个重要参数被进行参考。

目前常用的掌纹脊线距离估计方法为频域方法,但已有的频域方法对掌纹图像质量与完整性要求高,导致适用范围受限,且测量精度仍需进一步提高。 针对上述问题,提出一种基于注意力机制与残差网络的掌纹脊线距离估计方法,将手工标注获取的脊线距离值作为输入掌纹图像块的类别进行识别。 该方法以 VGG16 网络为基础,将 CBAM 注意力模块有机引入,增强网络对脊线结构信息的关注;并对损失函数进行改进,添加残差机制,来避免梯度消失问题。 所用数据集由手工标注完成,针对数据集中存在的样本分布不均衡问题,采用数据增强的方法削弱类别分布间的不平衡,并设计样本不均衡损失函数对模型进行优化。 实验结果表明, 该方法在该数据集上相比频域方法在识别准确率上提升了14.2% ,且相比部分经典深度学习方法在准确率上至少提升了 5. 1% 。

Abstract:

Palmprint ridge distance is one of the important texture attributes of palmprint ridge. It is often referred as an important　parameter in palmprint recognition algorithm and image quality evaluation method. At present,the commonly used method to estimatepalmprint ridge distance is frequency domain method,but the existing frequency domain method?

has high requirements on the quality andintegrity of palmprint image,resulting in limited application scope and further improvement of measurement accuracy. To solve theseproblems,a palmprint ridge distance estimation method based?

on attention mechanism and residual network is proposed. The ridgedistance value obtained by manual annotation is recognized as the category of the input palmprint image block. Based on VGG16network,CBAM attention module is introduced to enhance the network’s attention to ridge structure information. The loss function is improved and the residual mechanism is added to avoid the gradient disappearance. The data set is manually annotated. For the unbalanceddistribution of samples in the data set, we employ the method of data enhancement to weaken the imbalance between the classdistributions,and design the sample unbalanced loss function to optimize the model. Experimental results show that the proposed methodhas a 14. 2% improvement in recognition accuracy compared with the frequency domain methods on this dataset,and at least a 5. 1% improvement in recognition accuracy compared with some classical deep learning methods.

常用功能