基于多模态融合的人脸反欺骗技术

在先前的人脸反欺骗方法中大多使用手工提取的特征或者仅使用单一模态上的人 脸特征,并且很少注意到多通道色度的差异,因此得到的人脸反欺骗模型的鲁棒性较差以至于 无法有效地区分真假面孔。鉴于此,卷积神经网络(CNN)被用作特征提取器来代替手工特征的 提取,并且一种有效的多输入 CNN 模型被提出,以融合多种模态上的人脸特征以实现更具有 鲁棒性的人脸反欺骗。通过对人脸图像上的 2 个不同颜色特征(即 HSV 和 YCbCr)以及时间特征 进行联合建模,探索了人脸反欺骗的最佳鲁棒表示。在 REPLAY_ATTACK 和 CASIA-FASD 2 个基准数据集上进行的大量实验表明,该方法可实现最先进的性能。且在 REPLAY_ATTACK 上获得 0.23%的错误率(ERR)与 0.49%的半错误率(HTER)和在 CASIA-FASD 数据库上获得 1.76％的错误率与 3.05%的半错误率。     

基于条件对抗域泛化的人脸活体检测方法

针对现存的跨场景人脸活体检测模型泛化性能差、类间重叠等问题,提出了一种基于条件对抗域泛化的人脸活体检测方法。首先,该方法使用嵌入注意力机制的U-Net网络和ResNet-18编码器提取多个源域的特征,然后将提取的特征送入辅助分类器,并将特征编码器的输出和分类器预测的结果通过多线性映射的方法进行融合,再输入到域判别器中进行对抗训练,以实现特征和类层面对齐多个源域。其次,为了减少预测不准确的难迁移样本对域泛化造成的影响,采用了熵函数来控制样本的优先级,以提高域泛化的性能。此外,通过添加人脸深度图以进一步抓取活体与假体的区别特征,通过非对称三元组损失约束作为辅助监督,进一步提高类内紧凑性和类间区分性。在公开活体检测数据集上的对比实验验证了所提方法的有效性。     

基于区域分块和轻量级网络的人脸反欺骗方法

如何高效地辨别各种被攻击的人脸是人脸识别过程中迫切需要解决的问题。基于深度学习的人脸反欺骗方法在有着高性能的同时,也带来了庞大的参数量和计算量,使其无法部署在移动或嵌入式设备中。针对以上问题,提出了一种基于区域分块和轻量级网络的人脸反欺骗方法。首先,对训练样本进行随机区域分块;然后,设计了一种基于注意力机制的轻量级网络用于特征提取和图像分类;最后,为了提高测试准确率,对测试样本进行基于区域分块的数据扩增。实验结果表明,所提模型在CASIA-FASD和REPLAY-ATTACK数据集上达到了100%的准确率;在CASIA-SURF数据集的Depth模态上获得了99.49%的准确率和0.458 0%的平均分类错误率（ACER）,远优于ResNet、ShuffleNet等卷积神经网络,且该模型的参数量也仅有0.258 2 MB。在实际应用中,端到端的轻量级网络结构使所提模型更方便部署在移动设备上来进行实时的人脸反欺骗检测。     

基于内容风格增强和特征嵌入优化的人脸活体检测方法

针对现有人脸活体检测算法的特征表示不佳,以及在跨数据集上泛化性能较差等问题,提出了一种基于内容风格增强和特征嵌入优化的人脸活体检测方法。首先,使用ResNet-18编码器提取来自多个源域的通用特征,并经过不同注意力机制的两个自适应模块进行分离,增强全局内容特征与局部风格特征表征;其次,基于AdaIN算法将内容特征与风格特征进行有机融合,进一步提升特征表示,并将融合后的特征输入到特定的分类器和域判别器进行对抗训练;最后,采用平均负样本的半难样本三元组挖掘优化特征嵌入,可以兼顾类内聚集和类间排斥,更好地捕捉真实和伪造类别之间的界限。所提方法在四个基准数据集CASIA-FASD、REPLAY-ATTACK、MSU-MFSD 和 OULU-NPU上进行训练测试,分别达到了6.33%、12.05%、8.38%、10.59%的准确率,优于现有算法,表明所提方法能够显著提升人脸活体检测模型在跨数据集测试中的泛化性能。     

近红外高光谱图像数据预测技术

目的 受到传感器光谱响应范围的影响,可见光区域和近红外区域（400~2 500 nm）的高光谱数据通常使用不同的感光芯片进行成像,现有这一光谱区域典型的高光谱成像系统,如AVIRIS （airborne visible infrared imaging spectrometer）成像光谱仪,通常由多组感光芯片组成,整个成像系统成本和体积通常比较大,严重限制了该谱段高光谱探测技术的发展。为了能够扩展单感光芯片成像系统获得的高光谱图像的光谱范围,本文探索基于卷积神经网络的近红外光谱数据预测技术。方法 结合AVIRIS成像光谱仪的光谱配置,设计了基于残差学习的红外谱段图像预测网络,利用计算成像的方式从可见光范围的高光谱图像预测出近红外波段的光谱图像,并在典型的卫星高光谱遥感数据上进行红外光谱预测重构和基于重构的数据分类实验,以验证论文提出的红外光谱数据预测技术的可行性以及有效性。结果 本文设计的预测网络在Cuprite数据集上得到的预测近红外图像峰值信噪比为40.145 dB,结构相似度为0.996,光谱角为0.777 rad;在Salinas数据集上得到的预测近红外图像峰值信噪比为39.55 dB,结构相似性为0.997,光谱角为1.78 rad。在分类实验中,相比于只使用可见光图像,利用预测的近红外图像使得支持向量机（support vector machine,SVM）的准确率提升了0.6%,LeNet的准确率提升了1.1%。结论 基于AVIRIS传感器获取的两组典型卫星高光谱数据实验表明,本文提出的红外光谱数据预测技术不仅可基于计算成像的方式扩展可见光光谱成像系统的光谱成像范围,对于减小成像系统体积和质量具有重要意义,而且可有效提高可见光区域光谱图像数据在典型应用中的处理性能,对于提高高光谱数据处理精度提供新的技术支撑。     

Face anti-spoofing by identity masking using random walk patterns and outlier detection

Pattern Analysis and Applications - Existing architectures used in face anti-spoofing tend to deploy registered spatial measurements to generate feature vectors for spoof detection. This means that...     

CNN-based gender classification in near-infrared periocular images

Pattern Analysis and Applications - Periocular region has emerged as a key biometric trait with potential applications in the forensics domain. In this paper, we explore two convolutional neural...     

Illumination invariant face recognition using near-infrared images

Most current face recognition systems are designed for indoor, cooperative-user applications. However, even in thus-constrained applications, most existing systems, academic and commercial, are compromised in accuracy by changes in environmental illumination. In this paper, we present a novel solution for illumination invariant face recognition for indoor, cooperative-user applications. First, we present an active near infrared (NIR) imaging system that is able to produce face images of good condition regardless of visible lights in the environment. Second, we show that the resulting face images encode intrinsic information of the face, subject only to a monotonic transform in the gray tone; based on this, we use local binary pattern (LBP) features to compensate for the monotonic transform, thus deriving an illumination invariant face representation. Then, we present methods for face recognition using NIR images; statistical learning algorithms are used to extract most discriminative features from a large pool of invariant LBP features and construct a highly accurate face matching engine. Finally, we present a system that is able to achieve accurate and fast face recognition in practice, in which a method is provided to deal with specular reflections of active NIR lights on eyeglasses, a critical issue in active NIR image-based face recognition. Extensive, comparative results are provided to evaluate the imaging hardware, the face and eye detection algorithms, and the face recognition algorithms and systems, with respect to various factors, including illumination, eyeglasses, time lapse, and ethnic groups     

Face detection in color images

Human face detection plays an important role in applications such as video surveillance, human computer interface, face recognition, and face image database management. We propose a face detection algorithm for color images in the presence of varying lighting conditions as well as complex backgrounds. Based on a novel lighting compensation technique and a nonlinear color transformation, our method detects skin regions over the entire image and then generates face candidates based on the spatial arrangement of these skin patches. The algorithm constructs eye, mouth, and boundary maps for verifying each face candidate. Experimental results demonstrate successful face detection over a wide range of facial variations in color, position, scale, orientation, 3D pose, and expression in images from several photo collections (both indoors and outdoors)     

Face recognition in hyperspectral images

Hyperspectral cameras provide useful discriminants for human face recognition that cannot be obtained by other imaging methods. We examine the utility of using near-infrared hyperspectral images for the recognition of faces over a database of 200 subjects. The hyperspectral images were collected using a CCD camera equipped with a liquid crystal tunable filter to provide 31 bands over the near-infrared (0.7 /spl mu/m-1.0 /spl mu/m). Spectral measurements over the near-infrared allow the sensing of subsurface tissue structure which is significantly different from person to person, but relatively stable over time. The local spectral properties of human tissue are nearly invariant to face orientation and expression which allows hyperspectral discriminants to be used for recognition over a large range of poses and expressions. We describe a face recognition algorithm that exploits spectral measurements for multiple facial tissue types. We demonstrate experimentally that this algorithm can be used to recognize faces over time in the presence of changes in facial pose and expression.     

结合GAN的轻量级模糊车牌识别算法

目的 模糊车牌识别是车牌识别领域的难题,针对模糊车牌图像收集困难、车牌识别算法模型太大、不适用于移动或嵌入式设备等不足,本文提出了一种轻量级的模糊车牌识别方法,使用深度卷积生成对抗网络生成模糊车牌图像,用于解决现实场景中模糊车牌难以收集的问题,在提升算法识别准确性的同时提升了部署泛化能力。方法 该算法主要包含两部分,即基于优化卷积生成对抗网络的模糊车牌图像生成和基于深度可分离卷积网络与双向长短时记忆（long short-term memory,LSTM）的轻量级车牌识别。首先,使用Wasserstein距离优化卷积生成对抗网络的损失函数,提高生成车牌图像的多样性和稳定性;其次,在卷积循环神经网络的基础上,结合深度可分离卷积设计了一个轻量级的车牌识别模型,深度可分离卷积网络在减少识别算法计算量的同时,能对训练样本进行有效的特征学习,将特征图转换为特征序列后输入到双向LSTM网络中,进行序列学习与标注。结果 实验表明,增加生成对抗网络生成的车牌图像,能有效提高本文算法、传统车牌识别和基于深度学习的车牌识别方法的识别率,为进一步提高各类算法的识别率提供了一种可行方案。结合深度可分离卷积的轻量级车牌识别模型,识别率与基于标准循环卷积神经网络（convolutional recurrent neural network,CRNN）的车牌识别方法经本文生成图像提高后的识别率相当,但在模型的大小和识别速度上都优于标准的CRNN模型,本文算法的模型大小为45 MB,识别速度为12.5帧/s,标准CRNN模型大小是82 MB,识别速度只有7帧/s。结论 使用生成对抗网络生成图像,可有效解决模糊车牌图像样本不足的问题;结合深度可分离卷积的轻量级车牌识别模型,具有良好的识别准确性和较好的部署泛化能力。     

Face model fitting with learned displacement experts and multi-band images

Analyzing human faces is a traditional topic in computer vision research. For this task, model based approaches have been proven adequate to extract high-level information in many applications. However, they require a robust estimation of model parameters to work reliably. To tackle this challenge, we train displacement experts that serve as an update function on initial model parameter configurations. Unfortunately, building displacement experts that work robustly even in unconstrained environments is a non-trivial task. Therefore, we rely on a priori information about the structure of human faces by integrating an image representation that reflects the location of several facial components, so called “multi-band images”. By combining multi-band images and learned displacement experts, we propose a novel face model fitting approach. An evaluation on the “Labeled Faces In The Wild” database demonstrates, that this approach provides robust fitting results even in unconstrained environments.     

Face model fitting with learned displacement experts and multi-band images

In computer vision applications, models are often used to gain information about real-world objects. In order to determine model parameters that match the image content, displacement experts serve as an update function to refine initial model parameter estimations. However, building robust displacement experts is a non-trivial task, especially in unconstrained environments. Therefore, we provide the fitting algorithm not only with the original image but with a multi-band image representation that reflects the location of several facial components. To demonstrate its robustness in real-world scenarios, we integrate the Labeled Faces In The Wild database, which consists of images that have been taken outside lab environments.     

Useful features for human verification in near-infrared periocular images

The periocular region is the part of the face immediately surrounding the eye, and researchers have recently begun to investigate how to use the periocular region for recognition. Understanding how humans recognize faces helped computer vision researchers develop algorithms for face recognition. Likewise, understanding how humans analyze periocular images could benefit researchers developing algorithms for periocular recognition. We conducted two experiments to determine how humans analyze periocular images. In these experiments, we presented pairs of images and asked volunteers to determine whether the two images showed eyes from the same subject or from different subjects. In the first experiment, subjects were paired randomly to create different-subject queries. Our volunteers correctly determined the relationship between the two images in 92% of the queries. In the second experiment, we considered multiple factors in forming different-subject pairs; queries were formed from pairs of subjects with the same gender and race, and with similar eye color, makeup, eyelash length, and eye occlusion. In addition, we limited the amount of time volunteers could view a query pair. On this harder experiment, the correct verification rate was 79%. We asked volunteers to describe what features in the images were helpful to them in making their decisions. In both experiments, eyelashes were reported to be the most helpful feature.     

Ordinal regression with representative feature strengthening for face anti-spoofing

Neural Computing and Applications - Face anti-spoofing is a crucial link to ensure the security of face recognition. This paper proposes a novel face anti-spoofing method, which performs ordinal...     

Face recognition in low resolution thermal images

This paper proposes an accurate, rotation invariant, and fast approach for detection of facial features from thermal images. The proposed approach combines both appearance and geometric information to detect the facial features. A texture based detector is performed using Haar features and AdaBoost algorithm. Then the relation between these facial features is modeled using a complex Gaussian distribution, which is invariant to rotation. Experiments show that our proposed approach outperforms existing algorithms for facial features detection in thermal images. The proposed approach’s performance is illustrated in a face recognition framework, which is based on extracting a local signature around facial features. Also, the paper presents a comparative study for different signature techniques with different facial image resolutions. The results of this comparative study suggest the minimum facial image resolution in thermal images, which can be used in face recognition. The study also gives a guideline for choosing a good signature, which leads to the best recognition rate.     

基于多姿态人脸图像合成的识别方法研究

为了解决多姿态人脸识别问题,提出基于独立成分分析（ICA）进行正面人脸合成的新方法。首先利用ICA和PCA提取不同姿态人脸的特征子空间,然后利用通过训练得到的姿态转换矩阵合成其相对应的正面人脸图像,实验表明ICA人脸识别算法要优于PCA人脸识别算法,并在此基础上用小波对人脸图像进行预处理,据姿态转换矩阵得到的正面人脸特征系数直接进行分类比较,识别率得到了很大的提高。     

基于近红外与可见光双目视觉的活体人脸检测方法

针对人脸识别系统易受伪造攻击的问题,提出了一种基于近红外与可见光双目视觉的活体人脸检测方法。首先,采用近红外与可见光双目装置同步获取人脸图像,提取两图像的人脸特征点,利用双目关系实现特征点的匹配并获取其深度信息,再利用深度信息进行三维点云重建;然后,将全部人脸特征点划分为四个区域,计算各区域内人脸特征点在深度方向的平均方差;接着,选取人脸关键特征点,以鼻尖点为参照点,计算鼻尖点到人脸关键特征点之间的空间距离;最后,利用人脸特征点的深度值方差和空间距离来构造特征向量,使用支持向量机（SVM）实现活体人脸判断。实验结果表明,所提方法能够准确检测活体人脸以及有效抵御伪造人脸的攻击,在实验测试中达到99.0%的识别率,在准确性和鲁棒性上优于利用人脸特征点深度信息进行检测的同类算法。     

Detecting computer generated images based on local ternary count

Local binary patterns was used to distinguish the Photorealistic Computer Graphics and Photographic Images, however the dimension of the extracted features is too high. Accordingly, the Local Ternary Count based on the Local Ternary Patterns and the Local Binary Count was developed in this study. Furthermore, a novel algorithm is presented based on the Local Ternary Count to classify photorealistic Computer Graphics and Photographic images. The experiment results show that the proposed algorithm effectively reduces the dimension of the classification features and maintains a good classification performance.     

一种检测C&W对抗样本图像的盲取证算法

对抗样本图像能欺骗深度学习网络,亟待对抗样本防御机制以增强深度学习模型的安全性。C&W攻击是目前较热门的一种白盒攻击算法,它产生的对抗样本具有图像质量高、可转移、攻击性强、难防御等特点。本文以C&W攻击生成的对抗样本为研究对象,采用数字图像取证的思路,力图实现C&W对抗样本的检测,拒绝对抗样本输入深度学习网络。基于对抗样本中的对抗扰动易被破坏的假设,我们设计了基于FFDNet滤波器的检测算法。具体来说,FFDNet是一种基于深度卷积网络CNN的平滑滤波器,它能破坏对抗扰动,导致深度学习模型对对抗样本滤波前后的输出不一致。我们判断输出不一致的待测图像为C&W对抗样本。我们在ImageNet-1000图像库上针对经典的ResNet深度网络生成了6种C&W对抗样本。实验结果表明本文方法能较好地检测C&W对抗样本。相较于已有工作,本文方法不仅极大地降低了虚警率,而且提升了C&W对抗样本的检测准确率。