Improving representation-based classification for robust face recognition

The sparse representation classification (SRC) method proposed by Wright et al. is considered as the breakthrough of face recognition because of its good performance. Nevertheless it still cannot perfectly address the face recognition problem. The main reason for this is that variation of poses, facial expressions, and illuminations of the facial image can be rather severe and the number of available facial images are fewer than the dimensions of the facial image, so a certain linear combination of all the training samples is not able to fully represent the test sample. In this study, we proposed a novel framework to improve the representation-based classification (RBC). The framework first ran the sparse representation algorithm and determined the unavoidable deviation between the test sample and optimal linear combination of all the training samples in order to represent it. It then exploited the deviation and all the training samples to resolve the linear combination coefficients. Finally, the classification rule, the training samples, and the renewed linear combination coefficients were used to classify the test sample. Generally, the proposed framework can work for most RBC methods. From the viewpoint of regression analysis, the proposed framework has a solid theoretical soundness. Because it can, to an extent, identify the bias effect of the RBC method, it enables RBC to obtain more robust face recognition results. The experimental results on a variety of face databases demonstrated that the proposed framework can improve the collaborative representation classification, SRC, and improve the nearest neighbor classifier.     

An improved SRC method based on virtual samples for face recognition

The sparse representation classifier (SRC) performs classification by evaluating which class leads to the minimum representation error. However, in real world, the number of available training samples is limited due to noise interference, training samples cannot accurately represent the test sample linearly. Therefore, in this paper, we first produce virtual samples by exploiting original training samples at the aim of increasing the number of training samples. Then, we take the intra-class difference as data representation of partial noise, and utilize the intra-class differences and training samples simultaneously to represent the test sample in a linear way according to the theory of SRC algorithm. Using weighted score level fusion, the respective representation scores of the virtual samples and the original training samples are fused together to obtain the final classification results. The experimental results on multiple face databases show that our proposed method has a very satisfactory classification performance.     

Integrating GWTM and BAT algorithm for face recognition in low-resolution images

In biometrics, face recognition is one of the important identification methods with various applications such as, video surveillance, defence, human/computer interactions and many more. The current face recognition systems perform well using the frontal images with high resolution. In contrast, the utilisation of low-resolution (LR) images degrades the performance of face recognition systems. Hence, this paper integrates the Gabor filter?+?wavelet?+?texture (GWTM) operator and the BAT algorithm to increase the performance, while deploying the LR images. The proposed algorithm integrates the uniqueness of Gabor features, the robustness of local features and the wavelet features to handle the inter-person and intra-person variations. This paper utilises the spherical SVM classifier to enhance the recognition performance. Finally, the proposed GWTM operator is compared with other existing algorithms such as, GOM, LBP and LGP based on the parameters of accuracy, FAR and FRR. The proposed GWTM operator attains the highest accuracy of 95% and a minimum FAR of 5%. The results prove that the proposed GWTM yields a performance improvement of 5, 3, 4 and 15% over the GOM, LBP, LGP and GWTM, respectively, in the absence of the BAT algorithm.     

Combination of direct matching and collaborative representation for face recognition

Difficulties associated with the use of Buchdahl's retardation coefficients in image assessment are examined. It is shown that, by a series of approximations and corresponding transformations, the set of coordinates of transmitted rays from any object point can be expressed as a circular region perpendicular to the optical axis. Furthermore, it is shown that, under these transformations, the form of the retardation expansion remains constant and only the coefficients need be altered. These changes are independent of the field angle, but depend on the f-number of the system. The coefficients thus derived are field-independent in contrast to those specified by most authors. Expressions for the coefficients under each of the transformations introduced are presented. Also a brief discussion of the convergence of the retardation expansion is presented and the results indicate that the above approximations are sound over the region of convergence of the truncated aberration expansion of order eight.     

A supervised dimensionality reduction method-based sparse representation for face recognition

The sparse representation-based classification (SRC) method is a powerful tool to present high-dimensionality data and its superiority in many fields, especially in face recognition application has been proved. With sparsity appropriately harnessed, the SRC can solve face classification problems caused by varying expression, illumination as well as occlusion and disguise. However, face images as high-dimensionality data are usually noisy and the dimensionality is always larger than the number of training sample in real-world applications, which bring a disadvantage for the performance of SRC. Therefore, it is beneficial to perform dimensionality reduction (DR) before utilizing the SRC method. But most prevalent DR methods have no direct connection to SRC. In this paper, we proposed a supervised DR algorithm which suits SRC well and improves the discriminating ability in the low-dimensionality space. The proposed method utilizes the fisher discriminant criterion and low-dimensionality reconstructive restriction to extract the discriminating structure of data. The extensive experiments on public face databases verified the effectiveness of the supervised DR with the model of sparse representation.     

基于聚类和协同表示的超分辨率重建

图像超分辨率重建是利用单幅或多幅降质的低分辨率图像重建得到高分辨率图像,以提高图像的视觉效果并获得更多可用的信息。本文提出结合图像特征聚类和协同表示的超分辨率重建方法。在训练阶段根据图像的特征信息对图像样本进行聚类并利用图像特征的差异性训练不同的字典,克服了传统训练单个字典方法对图像特征表示不足的缺点。而且利用协同表示方法求得不同聚类的高、低分辨率图像样本之间的映射矩阵,提高了图像重建速度。实验表明,本文方法与其他方法相比,不仅提高了重建图像的PSNR和SSIM指标,而且改善了视觉效果。     

卫星遥感图像中类圆形油库的自动识别方法

油库是典型军事目标之一,对其识别是卫星图像判读的一项重要内容,传统的方法是通过判读员进行人工判读,工作量非常大是其缺点之一。为了克服这一缺点,本文提出了一种类圆形油库的自动识别方法。首先利用Kapur熵法对图像进行阈值分割,得到二值图像：然后对二值图像中的白像素进行最近邻聚类形成团块,并计算其面积以及体态比和矩形度等形状参数;最后利用油库近似圆形和成群分布的特点对油库群进行识别和定位。实验结果表明该方法对于高分辨率卫星遥感图像中的类圆形油库的识别是很有效的。     

Determination of candidate subjects for better recognition of faces

In order to improve the accuracy of face recognition and to solve the problem of various poses, we present an improved collaborative representation classification (CRC) algorithm using original training samples and the corresponding mirror images. First, the mirror images are generated from the original training samples. Second, both original training samples and their mirror images are simultaneously used to represent the test sample via improved collaborative representation. Then, some classes which are “close” to the test sample are coarsely selected as candidate classes. At last, the candidate classes are used to represent the test sample again, and then the class most similar to the test sample can be determined finely. The experimental results show our proposed algorithm has more robustness than the original CRC algorithm and can effectively improve the accuracy of face recognition.     

采用ASPCM模型进行姿势鲁棒性人脸识别

提出ASPCM模型,并将其用于不同姿势下的人脸识别。对人脸图像的形状表示和纹理表示进行主成分分析,建立形状模型和纹理模型;以形状参数、纹理参数和姿势参数间的转换确定人脸图像与头部角度间的映射关系;使用精确性和概括性两个标准衡量ASPCM模型的分解性能和合成性能;根据平均纹理相似度判断输入图像与模型视图间的相似程度。实验表明,该模型分解性能的精确性误差和概括性误差均在1.85°以内;合成性能的这两种误差均在1.1个像素以内;精确性和概括性的平均纹理相似度均在95.8%以上;当头部转动角度在25°以内时,该模型的识别率达到100%。     

人脸识别的一种新的特征提取方法

针对人脸识别中判别特征的提取问题,本文提出了一种新的人脸识别算法—扩展保局投影（ELPP）。普通保局投影（LPP）在构建权图时侧重保持样本的局部结构,属于无监督学习算法。扩展保局投影在保局投影的基础上进行扩展,通过引入可调因子,在保持人脸图像局部流形结构的同时考虑样本的类别信息,从而充分提取样本的判别特征。本文采用最小近邻分类器估算识别率。在Yale人脸库以及AT＆T人脸库的测试结果表明,在姿态、光照、表情、训练样本数目变化的情况下,ELPP都具有较好的识别率。     

人脸图像中人眼的检测与定位

利用人脸几何特征和图像分割原理，提出了一种在有背景的灰度和彩色人脸图像中自动检测与定位人眼的新算法。首先，基于人脸器官几何分布先验知识建立人眼位置判定准则；其次对人眼的分割阈值范围进行粗估计；然后采用分割阈值递增法，并结合人眼位置判定准则判定分割图像中双眼黑块是否出现；最后利用二雏相关系数作为对称相似性测度，检验检测到的双眼的真实性。为了避免图像背景对人眼检测的干扰，还运用了肤色分割原理来缩小检测人眼的搜索区域，从而进一步提高人眼定位的准确性。实验验证表明，所提出的人眼检测与定位方法在速度和准确性方面具有良好的性能。     

分块稀疏表示和2DPCA的人脸识别方法

稀疏表示提出了一种分块稀疏表示和二维主成分分析(2DPCA)的人脸识别方法.该方法应用了逐像素分块的与2DPCA技术相结合的方式,充分地考虑了图像中相邻的多个像素间的相关性.实验结果表明,其中提出的新算法具有可行性以及在识别精度上的优越性.进一步的研究还表明,所提出的分块识别的方法较之于以往传统算法在存在位置偏移、单色遮挡问题的人脸图像误判率上也有显著降低.     

利用人眼分形码距离进行快速人脸识别

在利用人脸分形码距离进行识别时,需要大量的时间对人脸库中每张人脸图像进行迭代与距离运算.为克服这一缺点,本文提出了用水平方向高频子带来定位眼睛并将其从人脸中抽取出来,进一步提出了基于人眼分形码距离的人脸快速识别算法.利用该算法,可去掉大部分人眼分形码距离较大的图像,从识别时间复杂性分析,本文算法所需时间主要与人眼大小以及用于最后识别的图像数目有关.在ORL和YALE两个人脸库上的实验结果表明,与本征脸方法和直接利用人脸分形码距离方法比较,在用于最后识别的图像数目占人脸库中人脸总数的20%左右时,本文算法可使平均识别率保持在约90%,与其它方法基本持平.     

神经网络在二维图像识别中的应用

本文提出了一种基于神经网络的二维图像识别技术。选取一组机械零件的二维图像,对每张图像进行放缩和旋转变换,并分析、提取对应图像的nmi特征和7个不变矩特征作为BP网络的输入样本,图像的二进制编号为输出样本构建BP神经网络。并对网络进行抗干扰训练,使网络对理想输入及带噪声的输入均有较好的识别率。实验证明该网络具有一定的工程实用性。     

Near-infrared and visible light image fusion algorithm for face recognition

In order to improve face recognition accuracy, we present a simple near-infrared (NIR) and visible light (VL) image fusion algorithm based on two-dimensional linear discriminant analysis (2DLDA). We first use two such schemes to extract two classes of face discriminant features of each of NIR and VL images separately. Then the two classes of features of each kind of images are fused using the matching score fusion method. At last, a simple NIR and VL image fusion approach is exploited to combine the scores of NIR and VL images and to obtain the classification result. The experimental results show that the proposed NIR and VL image fusion approach can effectively improve the accuracy of face recognition.     

采用改进的基于余量的算法实现人脸识别中最佳Gabor特征的选择

基于2DGabor变换的人脸特征描述已经受到了很多人的关注。然而现有的Gabor特征维数较高,而且具有冗余性,因此选择最佳的Gabor特征用于人脸识别显得尤为的重要。利用最大余量原理的特征选择算法在目前的机器学习研究中已经占据了重要的地位。本文在基于余量的迭代搜索法(Simba)的基础上,引入了一种新的选择算法:基于余量的共轭梯度法(Cgmba),它只需较少次迭代就可以找到最佳解。我们在IMM人脸库上进行了实验,实验结果表明:尽管只使用了一半不到的特征,但Cgmba和Simba的识别率却分别提高了3.75和1.25个百分点,同时也证实了我们提出的Cgmba明显优于Simba。最后我们对Cgmba选择的Gabor特征的分布情况进行了分析,可以看出较大尺度的特征相对于较小尺度的特征对于分辩人脸的细微差别具有同等的重要性,而且在垂直,135°方向的特征具有更强的分辩能力。     

改进的基于人眼结构特征的虹膜识别方法

本文提出了一种改进的基于人眼结构特征的虹膜识别方法,基本思想是:首先,通过图像预处理,确定虹膜的位置和大小,将环形虹膜图像展开成矩形并进行归一化与图像增强;其次,利用局部灰度极小值的方法寻找有效虹膜区域内的特征点,依据纹理长度和方向信息去掉伪特征点,得到虹膜二进制编码;最后根据匹配准则进行识别.大量实验表明,该方法识别率高,识别时间短.     

自适应的彩色图像光照补偿新方法

环境光照变化、光照分布不均匀会影响彩色图像中人脸的正确检测。在融合了不同的基于空域的解决方案的基础上,提出了一种自适应的彩色图像光照补偿新方法。分别对光照过亮、过暗以及中间灰度区域进行自适应的处理。对亮度值最小、最大的5%的像素,如果这些像素的数目足够多(本文大于100),在变换后分别压缩为0和255;用对数函数做非线性变换函数来修正中间灰度区域。在人脸PIE数据库上对光照不均的彩色图像进行了实验,验证了该方法能对人脸检测中的光照进行有效的补偿。     

一种基于三维模型投影的人脸识别新方法

研究了使用三维人脸模型进行不同姿势下的人脸识别问题,提出了一种三维建模二维识别的人脸识别算法,首先使用该方法将三维模型向不同方向投影,进而将不同姿势的二维图像与不同方向的投影结果相匹配,进行人脸识别。研究了使用Minolta Vivid 910进行数据获取,创建三维模型的方法和过程。实验结果表明,在进行不同姿势的人脸识别时,该方法的识别速度快于三维可变形模型方法,识别率远优于使用二维正面图像作为模板的人脸识别方法。