Kernel-based improved discriminant analysis and its application to face recognition

Kernel discriminant analysis (KDA) is a widely used tool in feature extraction community. However, for high-dimensional multi-class tasks such as face recognition, traditional KDA algorithms have the limitation that the Fisher criterion is nonoptimal with respect to classification rate. Moreover, they suffer from the small sample size problem. This paper presents a variant of KDA called kernel-based improved discriminant analysis (KIDA), which can effectively deal with the above two problems. In the proposed framework, origin samples are projected firstly into a feature space by an implicit nonlinear mapping. After reconstructing between-class scatter matrix in the feature space by weighted schemes, the kernel method is used to obtain a modified Fisher criterion directly related to classification error. Finally, simultaneous diagonalization technique is employed to find lower-dimensional nonlinear features with significant discriminant power. Experiments on face recognition task show that the proposed method is superior to the traditional KDA and LDA.     

Face recognition using a kernel fractional-step discriminant analysis algorithm

Feature extraction is among the most important problems in face recognition systems. In this paper, we propose an enhanced kernel discriminant analysis (KDA) algorithm called kernel fractional-step discriminant analysis (KFDA) for nonlinear feature extraction and dimensionality reduction. Not only can this new algorithm, like other kernel methods, deal with nonlinearity required for many face recognition tasks, it can also outperform traditional KDA algorithms in resisting the adverse effects due to outlier classes. Moreover, to further strengthen the overall performance of KDA algorithms for face recognition, we propose two new kernel functions: cosine fractional-power polynomial kernel and non-normal Gaussian RBF kernel. We perform extensive comparative studies based on the YaleB and FERET face databases. Experimental results show that our KFDA algorithm outperforms traditional kernel principal component analysis (KPCA) and KDA algorithms. Moreover, further improvement can be obtained when the two new kernel functions are used.     

基于散度差准则的隐空间特征抽取方法

本文提出了一种新的非线性特征抽取方法——基于散度差准则的隐空间特征抽取方法。该方法的主要思想就是首先利用一核函数将原始输入空间非线性变换到隐空间，然后，在该隐空间中，利用类间离散度与类内离散度之差作为鉴别准则进行特征抽取。与现有的核特征抽取方法不同，该方法不需要核函数满足Mercer定理，从而增加了核函数的选择范围。更为重要的是，由于采用了散度差作为鉴别准则，从根本上避免了传统的Fisher线性鉴别分析所遇到的小样本问题。在ORL人脸数据库和AR标准人脸库上的试验结果验证了本文方法的有效性。     

一种基于2DLPP和2DLDA的人脸识别方法研究

局部保持投影（locality preserving projection,LPP）和线性鉴别分析（linear discrimin antanalysis,LDA）是两种有效的一维特征提取方法,广泛应用于人脸识别领域。但采用一维特征提取方法时会存在列向量化时样本的结构信息被破坏和样本在提取特征时必须对协方差矩阵进行特征分解,对于高维小样本的问题很容易出现协方差矩阵奇异的问题。文中提出将二维局部保持投影（2DLPP）和二维线性鉴别分析（2DLDA）这两种方法在特征层进行融合并应用在人脸识别。基于人脸库AR上的实验表明,该方法比传统的IJPP和LDA识别性能更高,因此可作为一种新的人脸识别方法。     

人脸识别技术的研究

对人脸识别的几个关键技术进行了深入研究，提出了一种快速的基于眼睛像素特征的人脸检测方法：一种有效的基于SVD分解的特征提取方法和一种基于特征差别的SVM人脸识别方法．改进的基于SVD分解的特征提取方法能在一定程度上削弱光照和表情的影响，从而更好地抽取人脸的差别特征．基于特征差别的SVM方法将人脸识别这一典型的多分类问题构造成适合SVM处理的二分类问题，克服了传统SVM方法在解决多分类问题上的缺陷．实验表明该人脸检测方法有较高的正确检测率，提出的特征提取方法能有效地减弱光照和表情对人脸特征的负面影响，使得识别率有较大提高，基于特征差别的SVM方法有更好的概括能力和更高的正确识别率．     

适用于小样本问题的具有类内保持的正交特征提取算法

在人脸识别中, 具有正交性的特征提取算法是一类有效的特征提取算法, 但受到小样本问题的制约. 本文在正交判别保局投影的基础上, 提出了一种适用于小样本问题的具有类内保持的正交特征提取算法. 算法根据同类样本之间的空间结构信息, 重新定义了类内散度矩阵与类间散度矩阵, 进而给出了一个新的目标函数. 然而新的目标函数对于人脸识别问题, 同样存在着小样本问题. 为此本文将原始数据空间降到一个低维的子空间, 从而避免了总体散度矩阵奇异, 并在理论上证明了在该子空间中求解判别矢量集, 等价于在原空间中求解判别矢量集. 人脸库上的实验结果表明本文算法的有效性.     

基于多姿态特征融合生成对抗网络的人脸校正方法

针对人脸校正中单幅图像难以解决大姿态侧脸的问题,提出一种基于多姿态特征融合生成对抗网络（MFFGAN）的人脸校正方法,利用多幅不同姿态侧脸之间的相关信息来进行人脸校正,并采用对抗机制对网络参数进行调整。该方法设计了一种新的网络,包括由多姿态特征提取、多姿态特征融合、正脸合成三个模块组成的生成器,以及用于对抗训练的判别器。多姿态特征提取模块利用多个卷积层提取侧脸图像的多姿态特征;多姿态特征融合模块将多姿态特征融合成包含多姿态侧脸信息的融合特征;而正脸合成模块在进行姿态校正的过程中加入融合特征,通过探索多姿态侧脸图像之间的特征依赖关系来获取相关信息与全局结构,可以有效提高校正结果。实验结果表明,与现有基于深度学习的人脸校正方法相比,所提方法恢复出的正脸图像不仅轮廓清晰,而且从两幅侧脸中恢复出的正脸图像的识别率平均提高了1.9个百分点,并且输入侧脸图像越多,恢复出的正脸图像的识别率越高,表明所提方法可以有效融合多姿态特征来恢复出轮廓清晰的正脸图像。     

Multi-directional two-dimensional PCA with matching score level fusion for face recognition

The traditional matrix-based feature extraction methods that have been widely used in face recognition essentially work on the facial image matrixes only in one or two directions. For example, 2DPCA can be seen as the row-based PCA and only reflects the information in each row, and some structure information cannot be uncovered by it. In this paper, we propose the directional 2DPCA that can extract features from the matrixes in any direction. To effectively use all the features extracted by the D2DPCA, we combine a bank of D2DPCA performed in different directions to develop a matching score level fusion method named multi-directional 2DPCA for face recognition. The results of experiments on AR and FERET datasets show that the proposed method can obtain a higher accuracy than the previous matrix-based feature extraction methods.     

基于多姿态特征融合生成对抗网络的人脸校正方法

针对人脸校正中单幅图像难以解决大姿态侧脸的问题，提出一种基于多姿态特征融合生成对抗网络（MFFGAN）的人脸校正方法，利用多幅不同姿态侧脸之间的相关信息来进行人脸校正，并采用对抗机制对网络参数进行调整。该方法设计了一种新的网络，包括由多姿态特征提取、多姿态特征融合、正脸合成三个模块组成的生成器，以及用于对抗训练的判别器。多姿态特征提取模块利用多个卷积层提取侧脸图像的多姿态特征；多姿态特征融合模块将多姿态特征融合成包含多姿态侧脸信息的融合特征；而正脸合成模块在进行姿态校正的过程中加入融合特征，通过探索多姿态侧脸图像之间的特征依赖关系来获取相关信息与全局结构，可以有效提高校正结果。实验结果表明，与现有基于深度学习的人脸校正方法相比，所提方法恢复出的正脸图像不仅轮廓清晰，而且从两幅侧脸中恢复出的正脸图像的识别率平均提高了1.9个百分点，并且输入侧脸图像越多，恢复出的正脸图像的识别率越高，表明所提方法可以有效融合多姿态特征来恢复出轮廓清晰的正脸图像。     

隐空间中参数化直接鉴别分析及其应用

提出了一种新的非线性特征抽取方法——隐空间中参数化直接鉴别分析。其主要思想是利用一核函数将原始输入空间非线性变换到隐空间,针对在该隐空间中类内散布矩阵总是奇异等问题,利用参数化直接鉴别分析进行特征抽取。与现有的核特征抽取方法不同的是,该方法不需要核函数满足Mercer 定理,从而增加了核函数的选择范围。更为重要的是,由于在隐空间中采用了参数化直接鉴别分析,不仅保留了参数化直接鉴别分析的优点,而且有效地抽取了样本的非线性特征;在该方法中提出了一个更为合理的加权系数矩阵,提高了分类性能。在FERET人脸数据库子库上的实验结果验证了该方法的有效性。     

Aggregation of classifiers based on image transformations in biometric face recognition

In this study, we investigate the use of collective knowledge of independent classifiers (experts) in the area of face recognition. We formulate a hypothesis and provide compelling experimental evidence behind it that different image transformations can offer unique discriminatory information useful for face classification. We show that such discriminatory information can be combined in order to increase classification rates over those being produced by individual classifiers. In particular, we focus on contrast enhancement realized by histogram equalization and edge detection carried out with the use of the Sobel operator. We construct feature spaces emerging from linear and nonlinear methods of dimensionality reduction, namely Eigenfaces, Fisherfaces, kernel-PCA, and Isomap. Aggregation of classifiers is accomplished by majority voting and a Bayesian product rule. Extensive experimentation is conducted using the well-known FERET and YALE datasets.     

基于核的Fisher非线性最佳鉴别分析在人脸识别中的应用

抽取最佳鉴别特征是人脸识别中的重要一步。对小样本的高维人脸图像样本,由于各种抽取非线性鉴别特征的方法均存在各自的问题,为此提出了一种求解核的Fisher非线性最佳鉴别特征的新方法,该方法首先在特征空间用类间散度阵和类内散度阵作为Fisher准则,来得到最佳非线性鉴别特征,然后针对此方法存在的病态问题,进一步在类内散度阵的零空间中求解最佳非线性鉴别矢量。基于ORL人脸数据库的实验表明,该新方法抽取的非线性最佳鉴别特征明显优于Fisher线性鉴别分析(FLDA)的线性特征和广义鉴别分析(GDA)的非线性特征。     

基于多级纹理特征的深度信念网络人脸识别算法

为解决人脸特征提取过程中局部特征缺失的问题,借助局部二值模式(LBP)与方向梯度直方图(HOG)提出一种基于多级纹理特征融合的深度信念网络人脸识别算法。以提取局部纹理特征以及边缘纹理特征为出发点,对人脸图像进行三级纹理特征提取。使用MB-LBP提取初级纹理特征;在此基础上进行改进的CS-LBP图像特征提取作为二级纹理特征;使用HOG算子在二级纹理特征上完成三级纹理特征提取。将二级和三级纹理特征直方图顺序串联融合后输入到深度信念网络(DBN)逐层贪婪训练,优化网络参数,并用优化的网络在ORL、YELA人脸标准库中进行测试,识别率均在92%以上。该算法与传统算法(SVM、PCA)相比较拥有更好的人脸识别效果,同时也表明了局部纹理特征的改善为识别过程的特征提取提供强有力的保障,为人脸识别的进一步研究开拓新思路。     

一种基于块共同特征值的人脸识别方法

主成分分析与线性判别分析是人脸识别的重要识别方法,它们都通过求解特征值问题实现特征提取,但由于维数灾难会导致小样本和奇异性问题。提出了一种简单的人脸识别方法,无需进行奇异值分解,能有效地降低计算代价。首先将图像划分成块,然后计算多项式系数,得到友阵用于特征提取。基于两张不同图像的多项式系数友阵来计算对称阵。最后通过计算对称阵的零空间的零化度识别相似的人脸图像。为验证提出方法的有效性,在ORL、Yale和FERET人脸数据库上进行了实验。结果表明,该方法对于有较大姿态与光照变化的人脸识别具有较高的识别性能。     

基于卷积特征和贝叶斯分类器的人脸识别

为解决传统人脸识别算法特征提取困难的问题,提出了基于卷积特征和贝叶斯分类器的人脸识别方法,利用卷积神经网络提取人脸特征,通过主成分分析法对特征降维,最后利用贝叶斯分类器进行判别分类,在ORL（olivetti research laboratory）人脸库上进行实验,获得了99.00%的识别准确率。实验结果表明,卷积神经网络提取的人脸图像特征具有很强的辨识度,与PCA（principal component analysis）和贝叶斯分类器结合之后可有效提高人脸识别的准确率。     

融合全局与局部特征的贝叶斯人脸识别方法

针对人脸识别特征提取阶段中的数据降维方法往往难以兼顾保持全局与局部特征信息的问题，以及匹配识别阶段贝叶斯分类器中小样本问题，提出了一种融合全局与局部特征的贝叶斯人脸识别方法。该方法通过核主元分析提取出人脸数据的全局非线性特征，并在此基础上通过正交化局部敏感判别分析挖掘出人脸数据的局部流形结构信息，以达到提取出具有高判别力低维本质人脸特征的目的；采用一种最大信息量协方差选择的方法，来对协方差矩阵进行估算，以解决贝叶斯分类器设计中的小样本问题。在ORL、AR、 YALE、FLW人脸库上设计实验来进行验证。结果表明，提出的特征提取算法以及对贝叶斯分类器的改进取得了比较好的效果，通过对这两个阶段的优化，可以显著提升人脸识别的效果。     

结合零空间法和F-LDA的人脸识别算法

线性判别分析（LDA）是一种常用的线性特征提取方法。传统LDA应用于人脸识别时主要存在两个问题：1）小样本问题,即由于训练样本不足引起矩阵奇异; 2)优化准则函数并不直接与识别率相关。提出了一种新的能同时解决以上两个问题的基于LDA的人脸识别算法。首先,通过重新定义样本的类内散布矩阵和类间散布矩阵,提出了一种新的零空间法。然后把这种新的零空间法与F LDA（Fractional LDA）算法相结合,得到一种对人脸识别更有效的特征提取方法。实验结果表明,这种新算法具有较高的识别率。     

Face recognition using LDA-based algorithms

Low-dimensional feature representation with enhanced discriminatory power is of paramount importance to face recognition (FR) systems. Most of traditional linear discriminant analysis (LDA)-based methods suffer from the disadvantage that their optimality criteria are not directly related to the classification ability of the obtained feature representation. Moreover, their classification accuracy is affected by the "small sample size" (SSS) problem which is often encountered in FR tasks. In this paper, we propose a new algorithm that deals with both of the shortcomings in an efficient and cost effective manner. The proposed method is compared, in terms of classification accuracy, to other commonly used FR methods on two face databases. Results indicate that the performance of the proposed method is overall superior to those of traditional FR approaches, such as the eigenfaces, fisherfaces, and D-LDA methods.     

Constructing Facial Identity Surfaces for Recognition

We present a novel approach to face recognition by constructing facial identity structures across views and over time, referred to as identity surfaces, in a Kernel Discriminant Analysis (KDA) feature space. This approach is aimed at addressing three challenging problems in face recognition: modelling faces across multiple views, extracting non-linear discriminatory features, and recognising faces over time. First, a multi-view face model is designed which can be automatically fitted to face images and sequences to extract the normalised facial texture patterns. This model is capable of dealing with faces with large pose variation. Second, KDA is developed to compute the most significant non-linear basis vectors with the intention of maximising the between-class variance and minimising the within-class variance. We applied KDA to the problem of multi-view face recognition, and a significant improvement has been achieved in reliability and accuracy. Third, identity surfaces are constructed in a pose-parameterised discriminatory feature space. Dynamic face recognition is then performed by matching the object trajectory computed from a video input and model trajectories constructed on the identity surfaces. These two types of trajectories encode the spatio-temporal dynamics of moving faces.     

Efficient and robust feature extraction by maximum margin criterion

In pattern recognition, feature extraction techniques are widely employed to reduce the dimensionality of data and to enhance the discriminatory information. Principal component analysis (PCA) and linear discriminant analysis (LDA) are the two most popular linear dimensionality reduction methods. However, PCA is not very effective for the extraction of the most discriminant features, and LDA is not stable due to the small sample size problem . In this paper, we propose some new (linear and nonlinear) feature extractors based on maximum margin criterion (MMC). Geometrically, feature extractors based on MMC maximize the (average) margin between classes after dimensionality reduction. It is shown that MMC can represent class separability better than PCA. As a connection to LDA, we may also derive LDA from MMC by incorporating some constraints. By using some other constraints, we establish a new linear feature extractor that does not suffer from the small sample size problem, which is known to cause serious stability problems for LDA. The kernelized (nonlinear) counterpart of this linear feature extractor is also established in the paper. Our extensive experiments demonstrate that the new feature extractors are effective, stable, and efficient.