图神经网络与CNN融合的虹膜特征编码方法  

An iris feature-encoding method by fusion of graph neural networks and convolutional neural networks

作　　者：孙金通 沈文忠 Sun Jintong;Shen Wenzhong(School of Electronic and Information Engineering,Shanghai University of Electric Power,Shanghai 201306,China)

机构地区：[1]上海电力大学电子与信息工程学院,上海201306

出　　处：《中国图象图形学报》2024年第9期2764-2779,共16页Journal of Image and Graphics

摘　　要：目的更具可解释性的虹膜特征编码方法一直是虹膜识别中的一个关键问题,且低质量虹膜样本识别比较困难,图神经网络的发展为此类虹膜图像特征编码带来了新思路。本文提出了一种图神经网络与卷积神经网络融合的虹膜特征编码网络IrisFusionNet。方法在骨干网络前添加一个像素级增强模块以消除输入图像不确定性,并使用双分支骨干网络提取虹膜微观与宏观融合特征。训练阶段使用一个独特的联合损失函数对网络参数进行优化;推理阶段使用融合特征匹配策略进行特征匹配。结果实验结果表明:使用IrisFusionNet训练得到的特征提取器在多个公开低质量虹膜数据集上进行测试分别得到了EER(equal error rate)和FAR@FRR=0.01%的最佳值0.27%与0.84%,并且将分离度DI(discriminating index)提升30%以上,识别准确性以及类聚性均远远领先于基于卷积神经网络和其他使用图神经网络模型的虹膜识别优秀算法。结论本文提出的IrisFusionNet应用于虹膜识别任务具有极高的可行性和优越性。Objective Iris recognition is a prevalent biometric feature in identity recognition technology owing to its inher⁃ent advantages,including stability,uniqueness,noncontact modality,and live-body authentication.The complete iris rec⁃ognition workflow comprises four main steps:iris image acquisition,image preprocessing,feature encoding,and feature matching.Feature encoding serves as the core component of iris recognition algorithms.The improvement in interpretable iris feature encoding methods have become a pivotal concern in the field of iris recognition.Moreover,the recognition of low-quality iris samples,which often relies on specific parameter-dependent feature encoders,results in a poor generaliza⁃tion performance.The graph structure represents a data form with an irregular topological arrangement.Graph neural net⁃works(GNNs)effectively update and aggregate features within such graph structures.The advancement of GNN led to the development of new approaches for feature encoding of these types of iris images.In this paper,a pioneering iris featurefusion encoding network called IrisFusionNet,which integrates GNN with a convolutional neural network(CNN),is pro⁃posed.This network eradicates the need to implement complex parameter tuning steps and exhibits excellent generalization performance across various iris datasets.Method In the backbone network,the previously inserted pixel-level enhance⁃ment module alleviates local uncertainty in the input image through median filtering.In addition,global uncertainty was mitigated via Gaussian normalization.A dual-branch backbone network was proposed,where the head of the backbone net⁃work comprised a shared stack of CONV modules,and the neck was divided into two branches.The primary branch con⁃structed a graph structure from an image using graph converter.We designed a hard graph attention network that introduces an efficient channel attention mechanism to aggregate and update features through utilization of edge-associated information within the graph struc

关 键 词：虹膜特征编码 图神经网络(GNN) 硬图注意力算子 特征融合 联合损失函数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]