Overview of the combination of biometric matchers

Biometric identity verification refers to technologies used to measure human physical or behavioral characteristics, which offer a radical alternative to passports, ID cards, driving licenses or PIN numbers in authentication. Since biometric systems present several limitations in terms of accuracy, universality, distinctiveness, acceptability, methods for combining biometric matchers have attracted increasing attention of researchers with the aim of improving the ability of systems to handle poor quality and incomplete data, achieving scalability to manage huge databases of users, ensuring interoperability, and protecting user privacy against attacks. The combination of biometric systems, also known as “biometric fusion”, can be classified into unimodal biometric if it is based on a single biometric trait and multimodal biometric if it uses several biometric traits for person authentication.The main goal of this study is to analyze different techniques of information fusion applied in the biometric field. This paper overviews several systems and architectures related to the combination of biometric systems, both unimodal and multimodal, classifying them according to a given taxonomy. Moreover, we deal with the problem of biometric system evaluation, discussing both performance indicators and existing benchmarks.As a case study about the combination of biometric matchers, we present an experimental comparison of many different approaches of fusion of matchers at score level, carried out on three very different benchmark databases of scores. Our experiments show that the most valuable performance is obtained by mixed approaches, based on the fusion of scores. The source code of all the method implemented for this research is freely available for future comparisons¹.After a detailed analysis of pros and cons of several existing approaches for the combination of biometric matchers and after an experimental evaluation of some of them, we draw our conclusion and suggest some future directions of research, hoping that this work could be a useful start point for newer research.     

Likelihood ratio-based biometric score fusion

Multibiometric systems fuse information from different sources to compensate for the limitations in performance of individual matchers. We propose a framework for the optimal combination of match scores that is based on the likelihood ratio test. The distributions of genuine and impostor match scores are modeled as finite Gaussian mixture model. The proposed fusion approach is general in its ability to handle 1) discrete values in biometric match score distributions, 2) arbitrary scales and distributions of match scores, 3) correlation between the scores of multiple matchers, and 4) sample quality of multiple biometric sources. Experiments on three multibiometric databases indicate that the proposed fusion framework achieves consistently high performance compared to commonly used score fusion techniques based on score transformation and classification.     

Score normalization in multimodal biometric systems

Multimodal biometric systems consolidate the evidence presented by multiple biometric sources and typically provide better recognition performance compared to systems based on a single biometric modality. Although information fusion in a multimodal system can be performed at various levels, integration at the matching score level is the most common approach due to the ease in accessing and combining the scores generated by different matchers. Since the matching scores output by the various modalities are heterogeneous, score normalization is needed to transform these scores into a common domain, prior to combining them. In this paper, we have studied the performance of different normalization techniques and fusion rules in the context of a multimodal biometric system based on the face, fingerprint and hand-geometry traits of a user. Experiments conducted on a database of 100 users indicate that the application of min–max, z-score, and tanh normalization schemes followed by a simple sum of scores fusion method results in better recognition performance compared to other methods. However, experiments also reveal that the min–max and z-score normalization techniques are sensitive to outliers in the data, highlighting the need for a robust and efficient normalization procedure like the tanh normalization. It was also observed that multimodal systems utilizing user-specific weights perform better compared to systems that assign the same set of weights to the multiple biometric traits of all users.     

一种基于核函数分割数据集的分类器组合算法

组合分类器通过在输入空间中依据一定的规则生成数据集来训练成员分类器。提出一种新的基于核函数的模糊隶属度方法用来分隔数据集,并依据数据集中样本的模糊隶属度将它们分为相对难分和相对易分的数据子集,根据两个数据子集的难易程度训练不同的分类器。并用得到的两类分类器作为成员分类器生成组合分类器。将该组合分类器应用到UCI的标准数据集,实验表明该方法比Bagging和AdaBoost算法具有更好的性能。     

多分类问题代价敏感AdaBoost算法

针对目前多分类代价敏感分类问题在转换成二分类代价敏感分类问题存在的代价合并问题, 研究并构造出了可直接应用于多分类问题的代价敏感AdaBoost算法.算法具有与连续AdaBoost算法 类似的流程和误差估计. 当代价完全相等时, 该算法就变成了一种新的多分类的连续AdaBoost算法, 算法能够确保训练错误率随着训练的分类器的个数增加而降低, 但不直接要求各个分类器相互独立条件, 或者说独立性条件可以通过算法规则来保证, 但现有多分类连续AdaBoost算法的推导必须要求各个分类器相互独立. 实验数据表明, 算法可以真正实现分类结果偏向错分代价较小的类, 特别当每一类被错分成其他类的代价不平衡但平均代价相等时, 目前已有的多分类代价敏感学习算法会失效, 但新方法仍然能 实现最小的错分代价. 研究方法为进一步研究集成学习算法提供了一种新的思路, 得到了一种易操作并近似满足分类错误率最小的多标签分类问题的AdaBoost算法.     

基于基分类器系数和多样性的改进AdaBoost算法

针对传统AdaBoost算法的基分类器线性组合效率低以及过适应的问题,提出了一种基于基分类器系数与多样性的改进算法——WD AdaBoost。首先,根据基分类器的错误率与样本权重的分布状态,给出新的基分类器系数求解方法,以提高基分类器的组合效率;其次,在基分类器的选择策略上,WD AdaBoost算法引入双误度量以增加基分类器间的多样性。在五个来自不同实际应用领域的数据集上,与传统AdaBoost算法相比,CeffAda算法使用新的基分类器系数求解方法使测试误差平均降低了1.2个百分点;同时,WD AdaBoost算法与WLDF_Ada、AD_Ada、sk_AdaBoost等算法相对比,具有更低的错误率。实验结果表明,WD AdaBoost算法能够更高效地集成基分类器,抵抗过拟合,并可以提高分类性能。     

Fusion of color spaces for ear authentication

In this work, we propose a local approach for 2D ear authentication based on an ensemble of matchers trained on different color spaces. This is the first work that proposes to exploit the powerful properties of color analysis for improving the performance of an ear matcher.The method described is based on the selection of color spaces from which a set of Gabor features are extracted. The selection is performed using the sequential forward floating selection where the fitness function is related to the optimization of the ear recognition performance. Finally, the matching step is performed by means of the combination by the sum rule of several 1-nearest neighbor classifiers constructed on different color components.The effectiveness of the proposed method is demonstrated using the Notre-Dame EAR data set. Particularly interesting are the results obtained by the new approach in terms of rank-1 (∼84%), rank-5 (∼93%) and area under the ROC curve (∼98.5%), which are better than those obtained by other state-of-the-art 2D ear matchers.     

基于AdaBoost的微博垃圾评论识别方法

针对微博上存在的大量垃圾评论,提出一种基于AdaBoost的微博垃圾评论识别方法。该方法首先提取表示微博评论的特征值向量,由8个特征值组成,然后通过AdaBoost算法在这些特征上训练出若干个比随机预测好的弱分类器,最后将得到的弱分类器加权集合成高精度的强分类器。从实际的热门新浪微博中提取评论数据集进行实验,结果表明所选取的8个特征是有效的,该方法对于微博垃圾评论的识别拥有较高的识别率。     

Ensemble tracking

We consider tracking as a binary classification problem, where an ensemble of weak classifiers is trained online to distinguish between the object and the background. The ensemble of weak classifiers is combined into a strong classifier using AdaBoost. The strong classifier is then used to label pixels in the next frame as either belonging to the object or the background, giving a confidence map. The peak of the map and, hence, the new position of the object, is found using mean shift. Temporal coherence is maintained by updating the ensemble with new weak classifiers that are trained online during tracking. We show a realization of this method and demonstrate it on several video sequences     

基于全信息相关度的动态多分类器融合

AdaB00st采用级联方法生成各基分类器,较好地体现了分类器之间的差异性和互补性.其存在的问题是,在迭代的后期,训练分类器越来越集中在某一小区域的样本上,生成的基分类器体现不同区域的分类特征.根据基分类器的全局分类性能得到固定的投票权重,不能体现基分类器在不同区域上的局部性能差别.因此,本文基于Ada-Boost融合方法,利用待测样本与各分类器的全信息相关度描述基分类器的局部分类性能,提出基于全信息相关度的动态多分类器融合方法,根据各分类器对待测样本的局部分类性能动态确定分类器组合和权重.仿真实验结果表明,该算法提高了融合分类性能.     

结合旋转森林和AdaBoost分类器的多标签文本分类方法

针对一些多标签文本分类算法没有考虑文本-术语相关性和准确率不高的问题,提出一种结合旋转森林和AdaBoost分类器的集成多标签文本分类方法。首先,通过旋转森林算法对样本集进行分割,通过特征变换将各样本子集映射到新的特征空间,形成多个具有较大差异性的新样本子集。然后,基于AdaBoost算法,在样本子集中通过多次迭代构建多个AdaBoost基分类器。最后,通过概率平均法融合多个基分类器的决策结果,以此做出最终标签预测。在4个基准数据集上的实验结果表明,该方法在平均精确度、覆盖率、排名损失、汉明损失和1-错误率方面都具有优越的性能。     

A supervised method to discriminate between impostors and genuine in biometry

In this paper, we describe a supervised technique that allows to develop a more robust biometric system with respect to those based directly on the similarities of the biometric matchers or on the similarities normalised by the unconstrained cohort normalisation.In order to discriminate between genuine and impostors a quadratic discriminant classifier is trained using four features: the similarities of the biometric matcher; the similarities of the biometric matcher after the unconstrained cohort normalisation (UCN); the average scores among the test pattern and the users that belong to the background model; the difference between the user-specific threshold and the user-independent threshold.The proposed technique is validated by extensive experiments carried out on several biometric datasets (palm, finger, 2D and 3D faces, and ear). The experimental results demonstrate that the capabilities provided by our supervised method can significantly improve the performance of a standard biometric matcher or the performance of the standard UCN.     

Optimal resampling and classifier prototype selection in classifier ensembles using genetic algorithms

Ensembles of classifiers that are trained on different parts of the input space provide good results in general. As a popular boosting technique, AdaBoost is an iterative and gradient based deterministic method used for this purpose where an exponential loss function is minimized. Bagging is a random search based ensemble creation technique where the training set of each classifier is arbitrarily selected. In this paper, a genetic algorithm based ensemble creation approach is proposed where both resampled training sets and classifier prototypes evolve so as to maximize the combined accuracy. The objective function based random search procedure of the resultant system guided by both ensemble accuracy and diversity can be considered to share the basic properties of bagging and boosting. Experimental results have shown that the proposed approach provides better combined accuracies using a fewer number of classifiers than AdaBoost.     

Particle swarm optimisation based AdaBoost for object detection

This paper proposes a new approach to using particle swarm optimisation (PSO) within an AdaBoost framework for object detection. Instead of using exhaustive search for finding good features to be used for constructing weak classifiers in AdaBoost, we propose two methods based on PSO. The first uses PSO to evolve and select good features only, and the weak classifiers use a simple decision stump. The second uses PSO for both selecting good features and evolving weak classifiers in parallel. These two methods are examined and compared on two challenging object detection tasks in images: detection of individual pasta pieces and detection of a face. The experimental results suggest that both approaches can successfully detect object positions and that using PSO for selecting good individual features and evolving associated weak classifiers in AdaBoost is more effective than for selecting features only. We also show that PSO can evolve and select meaningful features in the face detection task.     

基于BP-AdaBoost神经网络的多参数掌静脉图像质量评价法

手掌静脉纹识别技术作为新一代高精度的生物特征识别技术,被广泛用于个人身份鉴定领域.然而,其识别效果受限于图像的质量,低质量的图像往往造成识别准确度偏低,如何有效的对图像质量进行评价从而筛选出高质量的图像成为掌静脉识别技术中的一项重要研究内容.本文旨在解决这一问题,提出了一种基于BP-AdaBoost神经网络的多参数的掌静脉图像质量评价法.根据掌静脉图像质量特点,提出多个参数的评价指标(对比度(contrast)、信息熵(entropy)、清晰度(sharpness)和等效视数(enl)).利用BP网络优良的非线性拟合特点,以多个评价参数为网络输入,分类结果为网络输出,训练10个BP弱分类器;在此基础上利用AdaBoost算法得到最终的强分类器.实验结果显示,对比传统加权融合的评价分类方法,分类的结果准确度较高,系统具有具有良好的应用价值.     

一种新的AdaBoost视频跟踪算法

针对复杂场景中运动目标较难定位的问题,提出一种结合纹理和颜色特征的AdaBoost目标跟踪算法.首先在线训练一个弱分类器的集合区分目标和背景;然后,通过AdaBoost将集合中的各弱分类器组合成一个强分类器,用于标定下一帧中各像素的类别属性,并生成置信图;最后,在置信图中用Mean Shift算法定位目标的中心.实验结果表明,该算法在光照变化、目标自身发生形变和遮挡的情况下,能准确地对目标进行跟踪.     

Troika - An improved stacking schema for classification tasks

Stacking is a general ensemble method in which a number of base classifiers are combined using one meta-classifier which learns their outputs. Such an approach provides certain advantages: simplicity; performance that is similar to the best classifier; and the capability of combining classifiers induced by different inducers. The disadvantage of stacking is that on multiclass problems, stacking seems to perform worse than other meta-learning approaches. In this paper we present Troika, a new stacking method for improving ensemble classifiers. The new scheme is built from three layers of combining classifiers. The new method was tested on various datasets and the results indicate the superiority of the proposed method to other legacy ensemble schemes, Stacking and StackingC, especially when the classification task consists of more than two classes.     

基于AdaBoost的组合分类器在遥感影像分类中的应用

运用组合分类器的经典算法AdaBoost将多个弱分类器-神经网络分类器组合输出,并引入混合判别多分类器综合规则,有效提高疑难类别的分类精度,进而提高分类的总精度.最后以天津地区ASTER影像为例,介绍了基于AdaBoost的组合分类算法,并在此基础上实现了天津地区的土地利用分类.分类结果表明,组合分类器能有效提高单个分类器的分类精度,分类总精度由81.13%提高到93.32%.实验表明基于AdaBoost的组合分类是遥感图像分类的一种新的有效方法.     

AdaBoost算法在车牌字符识别中的应用

提出了一种基于AdaBoost的车牌字符自动识别算法。AdaBoost是一种构建准确分类器的学习算法，它将一族弱学习算法通过一定规则结合成为一个强学习算法，从而通过样本训练得到一个识别准确率理想的分类器，将之用于车牌字符识别，对车牌图像进行实验，对车牌字符样本进行特征提取，用特征来训练有效分类器，用MATLAB完成了对车牌照数字识别的模拟，结果证实此算法对车牌字符识别有一定准确性，具有良好的效果。     

融合AdaBoost和启发式特征搜索的人脸性别分类

提出一种基于AdaBoost的人脸性别分类方法,从一张低分辨率灰度人脸图像中辨认出一个人的性别。将启发式搜索算法融于AdaBoost算法框架中,从而发现新的可用于更好分类的特征。利用该方法进行人脸性别分类方面的实验,当使用少于500个像素比较时,正确识别率达到了93%以上,这与迄今已公布的最佳的分类器支持向量机（SVM）的正确识别率相当,但速度却快得多。