基于多尺度特征提取与多元回归分析的人脸识别

为提高人脸识别的正确率,提出了一种改进的特征提取及分类算法。首先采用Contour-let变换对人脸图像进行多尺度分解,然后由低频子带和各尺度各方向的高频子带得到人脸的特征值,并将它们组合成多尺度特征向量,再应用多元回归分析方法进行人脸识别。由于多尺度特征向量不仅反映了整幅图像的全局特征,还反映了图像各种尺度下的边缘、纹理等奇异特征,因此具有更多的鉴别信息;多元回归分析则充分考虑了同一总体的各样本间的强线性关系。在ORL人脸库上的实验显示人脸识别率达97．78％,优于其他的方法。     

一种抗姿态与表情变化的三维人脸识别方法

为了提高人脸在姿态和表情变化下的识别率,结合局部平面距离(DLP)对曲面局部凹凸性优良的判断能力,提出了一种采用人脸的等距不变表示形式来匹配的人脸识别方法。首先,对深度摄像头采集到的深度图像进行距离约束、位置约束、转换等操作,得到干净完整的三维人脸,利用三维人脸上每一点DLP值确定鼻尖点,利用聚类的思想确定鼻根点;其次,采用改进的快速推进算法计算人脸的测地距矩阵,设置阈值并切割出有效的人脸区域;最后,计算有效的人脸区域的高阶矩特征,作为人脸的特征向量进行匹配。实验结果表明,对于不同的数据库,本文算法的识别率接近97%;将本文算法与基于轮廓线特征的人脸识别算法以及基于Gabor特征的人脸识别算法进行比较,其识别率分别提高了14.1%和8.3%,同时有着较高的运算效率。     

基于小波分解系数的贝叶斯人脸识别方法

本文给出了贝叶斯人脸识别方法中匹配准则的多个近似表达式及一种实用的快速计算方法.在此基础上,利用反对称双正交小波变换的微分算子功能,提出了一种利用两幅人脸图象的小波变换系数差作为模式矢量的贝叶斯人脸识别方法,并利用AR人脸图象库进行了实验.实验结果表明本文方法与基于图象灰度的类似方法相比,识别率提高8%左右.此外本文方法也提供了一条在图象压缩数据域中实现人脸识别的可能途径.     

一种基于样本生成的人脸识别方法

人脸识别问题的特点包括样本的特征维数高和每个类别所包含的样本较少。设计有效的特征提取方法是解决人脸识别问题的关键要素之一。提出了在采用降采样获得特征的同时利用新的降采样方法多次对原图片进行降采样,生成多幅训练样本,进而缓解人脸识别中的小样本问题。实验结果证明所提出的方法能有效地提高分类器的精度。     

基于小波分解系数的贝叶斯人脸识别方法

本文给出了贝叶斯人脸识别方法中匹配准则的多个近似表达式及一种实用的快速计算方法,在此基础上,利用反对称双正交小波变换的微分算子功能,提出了一种利用两幅人脸图像的小波变换系数差作为模式矢量的贝叶斯人脸识别方法,并利用AR人脸图象库进行了实验,实验结果表明本文方法与基于图像灰度的类似方法相比,识别率提高8％左右,此外本文方法也提供了一条在图像压缩数据域中实现人脸识别的可能途径。     

低照度条件下超分辨率人脸图像采集系统设计与实现

在计算机技术高速发展的时代,多平台计算机视觉库随之产生。OpenCV作为一种开源代码的计算机视觉库,以可兼容多平台、接口广泛的特点被广泛运用各个领域。在低照度条件下,会出现光照环境差异过大或光线不足等情况,导致传统图像采集系统不能采集高质量的人脸图像,局限性较差。提出基于OpenCV在C 环境配置下运用三维人脸识别技术算法,设计一套低照度条件下超分辨率人脸图像采集系统。实验证明,该设计方案具有实时(对焦速度快)、快速(单张采集0.05秒)、准确(面部识别率99.3%)等特点,能够充分满足低照度条件下超分辨率人脸图像采集的需求。     

一种对姿态稳健的鼻尖点快速定位算法

提出了一种对姿态稳健的鼻尖点快速定位算法。在局部基准坐标(LRF)下计算顶点的平面距离能量,并设计了一种新的迭代筛选算法,计算得到候选点;计算候选点集中的每个顶点在人脸三维矢量场中的散度,将散度值最大的顶点作为鼻尖点。在FRGC v2.0和Bosphorus人脸库上对算法进行验证,在Bosphorus库上最终平均每张人脸定位仅耗时0.62s,在FRGC v2.0库上的定位准确率为95.6%。最后与当前其他算法进行对比,所提算法在速度和精度上均取得了较好的结果。实验结果证明所提算法不仅有望达到实时处理的要求,还具有较高的准确率,且对人脸姿态变化具有稳健性。     

提高PCA识别率的新算法

针对基于主元分析(PCA)的识别算法不能最优区分不同种类样本的缺点,提出了一种新的多主元分析(Multi-PCA)识别算法。该算法为每类样本构造单独特征空间,用各个空间的特征向量重建待识别样本。特征空间是基于某类样本图像的共性建立,因此重建该类样本图像时将得到较小重建误差,而重建其它类图像时的误差较大。可以根据重建误差的大小来识别样本图像,将待识别样本分类到具有最小重建误差的特征空间。在ORL、YALE人脸库和标牌压印字符库上的实验结果显示,Multi-PCA的识别率比PCA有较大提高。     

基于Lucas-Kanade算法的最大Gabor相似度大姿态人脸识别

在人脸识别科学研究和实际应用领域中,大角度姿态是影响人脸识别结果的主要因素之一,成为限制人脸识别技术进步的难点,而姿态的校正归一化是解决该问题的常用手段。首先通过加权的LK(Lucas-Kanade)算法得到侧脸块和对应正脸块的仿射变换参数,基于最大Gabor相似度寻找校正人脸姿态的最优参数。然后,以每一人脸块最优参数得到的平均Gabor相似度作为这一块人脸的识别权重,可以增加大姿态人脸识别的精度和稳健性。在FERET人脸数据库中进行了实验,当水平偏转角度为45°时,准确率达到97.3%,证明本文提出的以最大Gabor相似度作为加权LK算法参数提取的依据是有效的,得到的最优参数具有较好的光照无关性,而将平均Gabor相似度作为识别权重,有助于使算法的应用更加稳健和有效。     

基于连续小波变换的神经网络人脸识别研究

研究了基于连续小波变换的神经网络进行人脸识别的方法.介绍了小波分析的理论基础，详细讨论了根据小波变换系数的范数选取小波母函数的方法，根据小波脊线确定网络神经元个数的方法以及神经网络的初始化和参数训练方法.通过对人脸图像灰度的连续小波分析，神经网络的自组织自学习能力，调整连接权值和小波神经元的尺度、位移参数，完成人脸识别的任务.实验结果验证了该神经网络的识别性能明显优于用特征脸方法对相同人脸库进行的识别.     

A sparse representation method based on kernel and virtual samples for face recognition

To improve the classification accuracy of face recognition, a sparse representation method based on kernel and virtual samples is proposed in this paper. The proposed method has the following basic idea: first, it extends the training samples by copying the left side of the original training samples to the right side to form virtual training samples. Then the virtual training samples and the original training samples make up a new training set and we use a kernel-induced distance to determine M nearest neighbors of the test sample from the new training set. Second, it expresses the test sample as a linear combination of the selected M nearest training samples and finally exploits the determined linear combination to perform classification of the test sample. A large number of face recognition experiments on different face databases illustrate that the error ratios obtained by our method are always lower more or less than face recognition methods including the method mentioned in Xu and Zhu [21], the method proposed in Xu and Zhu [39], sparse representation method based on virtual samples (SRMVS), collaborative representation based classification with regularized least square (CRC_RLS), two-phase test sample sparse representation (TPTSSR), and the feature space-based representation method.     

A hierarchical structure with improved OMP sparse representation used with face recognition

With the rapid development of the face recognition technology, more and more optical products are applied in people's real life. The recognition accuracy can be improved by increasing the number of training samples, but the colossal training samples will result in the increase of computational complexity. In recent years, sparse representation method becomes a research hot spot on face recognition. In this paper we propose an energy constrain orthogonal matching pursuit (ECOMP) algorithm for sparse representation to select the few training samples and a hierarchical structure for face recognition. We filter the training samples with ECOMP algorithm and then we compute the weights by all selected training samples. At last we find the closest recovery sample to the test sample. Simultaneously the experimental results in AR, ORL and FERET database also show that our proposed method has better recognition performance than the LRC and SRC_OMP method.     

A novel virtual samples-based sparse representation method for face recognition

A limited training set usually limits the performance of face recognition in practice. Even sparse representation-based methods which outperform in face recognition cannot avoid such situation. In order to effectively improve recognition accuracy of sparse representation-based methods on a limited training set, a novel virtual samples-based sparse representation (VSSR) method for face recognition is proposed in this paper. In the proposed method, virtual training samples are constructed to enrich the size and diversity of a training set and a sparse representation-based method is used to classify test samples. Extensive experiments on different face databases confirm that VSSR is robust to illumination variations and works better than many representative representation-based face recognition methods.     

Modular Fisher Discriminant Sparse Representation for robust face recognition

This paper proposes a novel framework for robust face recognition based on sparse representation and discrimination ranking. This method consists of three stages. The first stage partitions each training sample into some overlapped modules and then computes each module's Fisher ratio, respectively. The second stage selects modules which have higher Fisher ratios to comprise a template to filter training and test images. The dictionary is constructed by the filtered training images. The third stage computes the sparse representation of filtered test sample on the dictionary to perform identification. The advantages of the proposed method are listed as follows: the first stage can preserve the local structure. The second stage removes the modules that have little contribution for classification. Then the method uses the retaining modules to classify the test sample by SRC which makes the method robust. Compared with the related methods, experimental results on benchmark face databases verify the advancement of the proposed method. The proposed method not only has a high accuracy but also can be clearly interpreted.     

A novel pose and illumination robust face recognition with a single training image per person algorithm

In the real-world application of face recognition system, owing to the difficulties of collecting samples or storage space of systems, only one sample image per person is stored in the system, which is so-called one sample per person problem. Moreover, pose and illumination have impact on recognition performance. We propose a novel pose and illumination robust algorithm for face recognition with a single training image per person to solve the above limitations. Experimental results show that the proposed algorithm is an efficient and practical approach for face recognition.     

基于可见光谱和卷积神经网络的贝类识别方法

目前卷积神经网络（CNN）在物体种类识别方面取得突破性进展。贝类作为农业经济的重要组成部分,种类繁多,特点复杂,大多贝类存在着相似度高,各类样本分布不均衡情况,以致CNN对贝类分类的准确率偏低。针对这一情况,提出了基于可见光谱和CNN的贝类识别方法,旨在提取更有效的贝类特征,从而提高贝类分类的准确率。首先,提出了一种包含输出熵度量和正交性度量的滤波器信息度量与特征选择方法,重新初始化修剪掉的滤波器并使其正交,捕获网络激活空间中的不同方向,使神经网络模型学习到更多有用的贝类特征信息,提升模型分类准确率;其次,提出了一种包含正则化项和焦点损失项的贝类分类目标函数,通过控制各类别样本对总损失的共享权重,来减少易分类样本的权重,以使模型注意力向预测不准的样本倾斜,均衡样本分布和样本分类难度,进一步提高贝类分类的准确率。贝类图像数据集由74类贝类组成,共11 803张图像。获取原始数据集后,对数据集图像进行水平翻转、垂直翻转、随机旋转、在[0, 30°]范围内旋转、在[0, 20%]范围内缩放和移动等数据增强操作,将图像数量从11 803张增加到119 964张。整个图像数据集按8∶1∶1的比例随机分为训练集95 947张图片、验证集11 996张图片和测试集12 021张图片。在建立贝类图像数据集的基础上进行了实验验证,达到了93.38%的分类准确率,将基准网络（Resnest）的准确率提高了1.18%,相较网络SN_Net和MutualNet,准确率分别提升了4.34%和0.85% ,并且训练时长为22 320 s,将基准网络（Resnest）的训练时长缩短了960 s,训练时长分别比SN_Net和MutualNet短3 180和2 460 s。实验结果证明了该方法的有效性。     

数据驱动模型的血液物种光谱检测技术

介绍一种基于光谱检测和数据驱动模型的非接触式血液物种识别技术。选取了4个物种(猴144,大鼠203,狗133,人169)共计649个血样作为原始样本。超连续谱激光光源的波长范围是450～2 400 nm。分别采集抗凝管盛装血液样本的后向散射可见光谱(294～1 160 nm)和十个不同空间位点的前向散射近红外光谱(1 021～1 757 nm),将十一条光谱数据顺序连接为一维数据作为每个样本的原始数据。利用主成分分析法对数据集进行特征信息提取,保留原始差异信息量的99.99%,同时将数据量压缩为原始数据量的1.5%,提高分类识别的运算效率。对不同数量的训练集和验证集进行训练预测实验表明,十折交叉验证的识别误差率随着样本数量的增加而降低,样本库规模的增大可以提高识别的精确度。由于数据驱动模型是基于机器学习算法的数据流处理模型,因而可以采用多种不同的分类算法实现。通过比较人工神经网络、支持向量机、偏最小二乘回归、多元线性回归、随机森林和朴素贝叶斯的识别效果可以发现,不同算法的识别效果具有类别差异性,即各个算法的正确识别率排序在不同的物种中是有差异的。因而实际应用中,在选择数据驱动模型时,除了需要考虑算法的整体识别率之外,当对部分类别的识别效果有额外要求时,还应该考虑算法本身的类别差异性。     

A novel method for recognizing face with partial occlusion via sparse representation

In this paper, we propose a novel method to recognize the partially occluded face images based on sparse representation. Generally, occlusions, such as glasses and scarf, fall on some random patch of image's pixels of test images, but which is known to be connected. In our method, all images are divided into several blocks and then an indicator based on linear regression technique is presented to determine whether a block is occluded. We utilize those uncontaminated blocks as the new feature of an image. Finally, the sparse representation-based classification (SRC) method serves as the classifier to recognize unknown faces. In the original work of SRC, the extended SRC (eSRC) scheme is presented to deal with occlusions, which has very heavy computational cost. The experimental results show that our method can achieve good recognition accuracy and has much lower computational cost than eSRC.     

Face image recognition via collaborative representation on selected training samples

In this paper, we present a collaborative representation-based classification on selected training samples (CRC_STS) for face image recognition. The CRC_STS uses a two stage scheme: The first stage is to select some most significant training samples from the original training set by using a multiple round of refining process. The second stage is to use collaborative representation classifier to perform classification on the selected training samples. Our method can be regarded as a sparse representation approach but without imposing l₁-norm constraint on representation coefficients. The experimental results on three well known face databases show that our method works very well.     

Integrating the original face images and “symmetrical faces” to perform face recognition

Using the original and ‘symmetrical face’ training samples to perform representation based face recognition was first proposed in [1]. It simultaneously used the original and ‘symmetrical face’ training samples to perform a two-step classification and achieved an outstanding classification result. However, in [1] the “symmetrical face” is devised only for one method. In this paper, we do some improvements on the basis of [1] and combine this “symmetrical faces” transformation with several representation based methods. We exploit all original training samples, left “symmetrical face” training samples and right “symmetrical face” training samples for classification and use the score fusion for ultimate face recognition. The symmetry of the face is first used to generate new samples, which is different from original face image but can really reflect some possible appearance of the face. It effectively overcomes the problem of non-sufficient training samples. The experimental results show that the proposed scheme can be used to improve a number of traditional representation based methods including those that are not presented in the paper.