Fingerprint classification using fast Fourier transform and nonlinear discriminant analysis

In this paper, we present a new approach for fingerprint classification based on discrete Fourier transform (DFT) and nonlinear discriminant analysis. Utilizing the DFT and directional filters, a reliable and efficient directional image is constructed from each fingerprint image, and then nonlinear discriminant analysis is applied to the constructed directional images, reducing the dimension dramatically and extracting the discriminant features. The proposed method explores the capability of DFT and directional filtering in dealing with low-quality images and the effectiveness of nonlinear feature extraction method in fingerprint classification. Experimental results demonstrates competitive performance compared with other published results.     

Fingerprint recognition using mel-frequency cepstral coefficients

This paper presents a new fingerprint recognition method based on mel-frequency cepstral coefficients (MFCCs). In this method, cepstral features are extracted from a group of fingerprint images, which are transformed first to 1-D signals by lexicographic ordering. MFCCs and polynomial shape coefficients are extracted from these 1-D signals or their transforms to generate a database of features, which can be used to train a neural network. The fingerprint recognition can be performed by extracting features from any new fingerprint image with the same method used in the training phase. These features are tested with the neural network. The different domains are tested and compared for efficient feature extraction from the lexicographically ordered 1-D signals. Experimental results show the success of the proposed cepstral method for fingerprint recognition at low as well as high signal to noise ratios (SNRs). Results also show that the discrete cosine transform (DCT) is the most appropriate domain for feature extraction.     

Robust hand image processing for biometric application

A new approach for both hand image segmentation and feature extraction is described. The main advantage of this approach, namely its robustness to low quality images, is illustrated through verification experiments with two public databases: one with scanned images from 50 subjects and another one with low-quality images acquired from 23 subjects, from a conventional webcam. In both cases, features are successfully extracted and good performances are obtained, in spite of image quality. Moreover, the main drawbacks of feature extraction in conventional algorithms are highlighted.     

基于方向滤波分割的指纹自动识别系统算法

为适应当前个人身份鉴定迫切需要，提出了一种基于方向滤波分割的指纹自动识别系统的完整方案。该系统主要分指纹预处理及识别两大部分，包括方向图的计算、平滑、方向滤波、二值化、去噪、特征提取、特征编码及识别的方法。利用该系统的预处理及特征编码和识别方法，指纹图象得到了较好的增强，而且，在识别过程中能克服一定范围内指纹旋转及平移因素的影响。本系统具有高效识别和很强的抗干扰能力，对于残缺或模糊的指纹图象也有很好的识别效果，有较高的使用价值。     

Robust watermarking of fingerprint images

This paper introduces two spatial methods in order to embed watermark data into fingerprint images, without corrupting their features. The first method inserts watermark data after feature extraction, thus preventing watermarking of regions used for fingerprint classification. The method utilizes an image adaptive strength adjustment technique which results in watermarks with low visibility. The second method introduces a feature adaptive watermarking technique for fingerprints, thus applicable before feature extraction. For both of the methods, decoding does not require original fingerprint image. Unlike most of the published spatial watermarking methods, the proposed methods provide high decoding accuracy for fingerprint images. High data hiding and decoding performance for color images is also observed.     

Learning fingerprint minutiae location and type

For simplicity of pattern recognition system design, a sequential approach consisting of sensing, feature extraction and classification/matching is conventionally adopted, where each stage transforms its input relatively independently. In practice, the interaction between these modules is limited. Some of the errors in this end-to-end sequential processing can be eliminated, especially for the feature extraction stage, by revisiting the input pattern. We propose a feedforward of the original grayscale image data to a feature (minutiae) verification stage in the context of a minutiae-based fingerprint verification system. This minutiae verification stage is based on reexamining the grayscale profile in a detected minutia's spatial neighborhood in the sensed image. We also show that a feature refinement (minutiae classification) stage that assigns one of two class labels to each detected minutia (ridge ending and ridge bifurcation) can improve the matching accuracy by ∼1% and when combined with the proposed minutiae verification stage, the matching accuracy can be improved by ∼3.2% on our fingerprint database.     

多印痕指纹识别的数据融合技术

由于指纹图像中存在非线性变形和噪声干扰,当前的指纹识别系统不能满足某些系统要求保持很低的误接受率(FAR)的同时达到较低误拒绝率(FRR)的要求。提高识别性能的一个有效的方法是融合多个模板和多个指纹的数据融合技术。该文研究的一种多印痕指纹数据融合技术能够改善系统性能。通过实验证明了这种数据融合技术能够有效提高系统性能,可以使系统在保持低误接受率的条件下降低误拒绝率。     

A pitfall in fingerprint bio-cryptographic key generation

The core of bio-cryptography lies in the stability of cryptographic keys generated from uncertain biometrics. It is essential to minimize every possible uncertainty during the biometric feature extraction process. In fingerprint feature extraction, it is perceived that pixel-level image rotation transformation is a lossless transformation process. In this paper, an investigation has been conducted on analyzing the underlying mechanisms of fingerprint image rotation processing and potential effect on the major features, mainly minutiae and singular point, of the rotation transformed fingerprint. Qualitative and quantitative analyses have been provided based on intensive experiments. It is observed that the information integrity of the original fingerprint image can be significantly compromised by image rotation transformation process, which can cause noticeable singular point change and produce a non-negligible number of fake minutiae. It is found that the quantization and interpolation process can change the fingerprint features significantly without affecting the visual image. Experiments show that up to 7% bio-cryptographic key bits can be affected due to this rotation transformation.     

指纹图像预处理和特征提取

根据指纹的固有规律，提出了一套较完整的指纹图像预处理和特征提取算法。改进了指纹脊线频率求取算法和二值图像去孔噪声算法，并提出了一种新的滤除指纹伪特征点的方法。通过对上百幅不同质量的指纹图像进行测试，获得了较好的效果。     

Two-class fingerprint matcher

We present a system for fingerprint verification that approaches the problem as a two-class pattern recognition problem. The features extracted by “FingerCode” are used to capture the ridge strength. This feature vector is then classified as genuine or impostor according to a novel approach to handle the fingerprint verification as a two-class problem. Moreover, we show that extracting the features from sub-images around the core permits to better represent the local information.     

一种基于矩阵补全的室内指纹定位算法

近年来室内定位技术引起了研究者的广泛关注,现有基于信号指纹的室内定位算法需要大量采集指纹数据,且在噪声干扰下易产生较大的定位误差。针对上述问题,提出了一种鲁棒的基于矩阵补全的室内指纹定位算法,其基于信号指纹矩阵的低秩特性,将噪声干扰下的指纹数据恢复问题建模为范数正则化矩阵补全问题;在此基础上引入L1范数和F范数以平滑野值噪声并提高算法的稳定性,最终通过交替方向乘子法和变量分裂技术进行有效求解。实验结果表明,该算法只需进行少量信号指纹数据采集即可较为完整地恢复出指纹库,在各种噪声场景下均能获得高于同类算法的定位精度。     

3D fingerprint reconstruction system using feature correspondences and prior estimated finger model

The paper studies a 3D fingerprint reconstruction technique based on multi-view touchless fingerprint images. This technique offers a solution for 3D fingerprint image generation and application when only multi-view 2D images are available. However, the difficulties and stresses of 3D fingerprint reconstruction are the establishment of feature correspondences based on 2D touchless fingerprint images and the estimation of the finger shape model. In this paper, several popular used features, such as scale invariant feature transformation (SIFT) feature, ridge feature and minutiae, are employed for correspondences establishment. To extract these fingerprint features accurately, an improved fingerprint enhancement method has been proposed by polishing orientation and ridge frequency maps according to the characteristics of 2D touchless fingerprint images. Therefore, correspondences can be established by adopting hierarchical fingerprint matching approaches. Through an analysis of 440 3D point cloud finger data (220 fingers, 2 pictures each) collected by a 3D scanning technique, i.e., the structured light illumination (SLI) method, the finger shape model is estimated. It is found that the binary quadratic function is more suitable for the finger shape model than the other mixed model tested in this paper. In our experiments, the reconstruction accuracy is illustrated by constructing a cylinder. Furthermore, results obtained from different fingerprint feature correspondences are analyzed and compared to show which features are more suitable for 3D fingerprint images generation.     

Modelling fingerprint ridge orientation using Legendre polynomials

The estimation of fingerprint ridge orientation is an essential step in every automatic fingerprint verification system. The importance of ridge orientation can be deflected from the fact that it is inevitably used for detecting, describing and matching fingerprint features such as minutiae and singular points. In this paper we propose a novel method for fingerprint ridge orientation modelling using Legendre polynomials. One of the main problems it addresses is smoothing orientation data while preserving details in high curvature areas, especially singular points. We show that singular points, which result in a discontinuous orientation field, can be modelled by the zero-poles of Legendre polynomials. The models parameters are obtained in a two staged optimization procedure. Another advantage of the proposed method is a very compact representation of the orientation field, using only 56 coefficients. We have carried out extensive experiments using a state-of-the-art fingerprint matcher and a singular point detector. Moreover, we compared the proposed method with other state-of-the-art fingerprint orientation estimation algorithms. We can report significant improvements in both singular point detection and matching rates.     

指纹识别系统中的图像预处理技术

根据指纹的特点,提出了一套有效的指纹识别预处理算法.包括前背景分离、基于空域和频域的图像增强、方向滤波、二值化、细化以及细化去噪处理.由于常用算法在指纹处理中存在缺陷,本文改进了增强算法和细化算法.实验证明,经过本算法处理后,减少了指纹图像中的伪结点和原有特征的丢失,有利于指纹特征提取.     

多尺度注意力融合和抗噪声的轻量点云人脸识别模型

在低质量点云人脸数据集上,判别性特征的提取和模型对噪声的鲁棒性是解决点云人脸识别问题 的关键。针对现有轻量点云人脸识别算法不能充分提取判别性特征和数据集中存在大量噪声而影响模型训练的问 题,设计轻量高效的网络模型,提出了基于多尺度注意力融合和抗噪声的自适应损失函数的点云人脸识别算法。 首先通过不同卷积模块获得不同感受野大小的特征图。然后进行多尺度的注意力特征提取,并使用高层的注意力 权重来引导低层注意力权重的生成,最后进行通道融合得到多尺度融合的特征,提升了模型捕获人脸细节特征的 能力。其次,根据低质量点云人脸图像的噪声信息特点,设计了一种新颖的抗噪声的自适应损失函数(anti-noise adaptive loss),以应对数据集大量噪声对模型训练过程中可能造成的负面影响,提升模型的鲁棒性和泛化能力。 在开源数据集 Lock3DFace 和本文提出的 KinectFaces 数据集上的实验结果表明,与当前的主流算法相比该算法模 型在低质量点云人脸识别任务中具有更好的识别效果。     

模糊几何和纹理特征在指纹细节点后处理中的应用

指纹识别一般基于指纹细节点匹配，当指纹图像质量较差时，细节点的可靠提取十分困难，通常会产生大量的虚假细节点．为提高细节点的精度，给出一种在原始灰度指纹图像上进行细节点后处理验证的方法．在每个自动提取出的细节点上取其在原始灰度指纹图像上的局部邻域，分析邻域图像的模糊几何特征和纹理特征，然后用MLP神经网络对提取出的局部邻域特征进行分类，实现细节点类型验证．实验结果表明：文中方法能有效地去除大量的虚假细节点，与其他方法相比具有较高精度．     

指纹纹线的提取和匹配

针对传统指纹细节点提取和匹配方法在处理低质量图像上的不足,首先提出了一种利用复滤波器准确定位指纹图像中心点的方法,克服了传统自动指纹识别中寻找指纹中心点不够精确的缺点。同时提出了一套基于轮廓的指纹特征提取和匹配算法,此算法利用指纹的总体结构特征,提高了匹配的正确率。实验结果表明,该算法具有相当高的识别率和较强的鲁棒性。     

Spoofing protection for fingerprint scanner by fusing ridge signal and valley noise

Biometric fingerprint scanners are positioned to provide improved security in a great span of applications from government to private. However, one highly publicized vulnerability is that it is possible to spoof a variety of fingerprint scanners using artificial fingers made from Play-Doh, gelatin and silicone molds. Therefore, it is necessary to offer protection for fingerprint systems against these threats. In this paper, an anti-spoofing detection method is proposed which is based on ridge signal and valley noise analysis, to quantify perspiration patterns along ridges in live subjects and noise patterns along valleys in spoofs. The signals representing gray level patterns along ridges and valleys are explored in spatial, frequency and wavelet domains. Based on these features, separation (live/spoof) is performed using standard pattern classification tools including classification trees and neural networks. We test this method on a larger dataset than previously considered which contains 644 live fingerprints (81 subjects with 2 fingers for an average of 4 sessions) and 570 spoof fingerprints (made from Play-Doh, gelatin and silicone molds in multiple sessions) collected from the Identix fingerprint scanner. Results show that the performance can reach 99.1% correct classification overall. The proposed anti-spoofing method is purely software based and integration of this method can provide protection for fingerprint scanners against gelatin, Play-Doh and silicone spoof fingers.     

基于边际Fisher深度自编码器的电台指纹特征提取*

针对在少量有标签样本条件下传统方法训练不充分而且难以准确提取通信电台指纹特征的问题,文中提出基于边际Fisher深度自编码器的电台指纹特征提取算法.以深度自编码器为基础,训练过程分为无监督预训练、基于边际Fisher映射的有监督训练两部分.首先挖掘海量无标签样本中包含的电台个体类别信息,用于深度自编码器最优参数训练.然后在有标签样本的辅助下对训练参数进行基于边际Fisher映射的有监督精校,提高指纹特征对同类型电台个体的鉴别能力.在多个通信电台数据集上进行的分类识别实验表明,文中算法能在小样本训练条件下有效表达同类型通信电台个体之间的差异.