Face anti-spoofing algorithm combined with CNN and brightness equalization

Face anti-spoofing is a relatively important part of the face recognition system,which has great significance for financial payment and access control systems.Aiming at the problems of unstable face alignment,complex lighting,and complex structure of face anti-spoofing detection network,a novel method is presented using a combination of convolutional neural network and brightness equalization.Firstly,multi-task convolutional neural network(MTCNN)based on the cascade of three convolutional neural networks(CNNs),P-net,R-net,and O-net are used to achieve accurate positioning of the face,and the detected face bounding box is cropped by a specified multiple,then brightness equalization is adopted to perform brightness compensation on different brightness areas of the face image.Finally,data features are extracted and classification is given by utilizing a 12-layer convolution neural network.Experiments of the proposed algorithm were carried out on CASIA-FASD.The results show that the classification accuracy is relatively high,and the half total error rate(HTER)reaches 1.02%.

A Survey on Face Anti-Spoofing Algorithms

The development of artificial intelligence makes the application of face recognition more and more extensive,which also leads to the security of face recognition technology increasingly prominent.How to design a face anti-spoofing method with high accuracy,strong generalization ability and meeting practical needs is the focus of current research.This paper introduces the research progress of face anti-spoofing algorithm,and divides the existing face anti-spoofing methods into two categories:methods based on manual feature expression and methods based on deep learning.Then,the typical algorithms included in them are classified twice,and the basic ideas,advantages and disadvantages of these algorithms are analyzed.Finally,the methods of face anti-spoofing are summarized,and the existing problems and future prospects are expounded.

Research on Face Anti-Spoofing Algorithm Based on Image Fusion

Along with the rapid development of biometric authentication technology,face recognition has been commercially used in many industries in recent years.However,it cannot be ignored that face recognition-based authentication techniques can be easily spoofed using various types of attacks such photographs,videos or forged 3D masks.In order to solve this problem,this work proposed a face anti-fraud algorithm based on the fusion of thermal infrared images and visible light images.The normal temperature distribution of the human face is stable and characteristic,and the important physiological information of the human body can be observed by the infrared thermal images.Therefore,based on the thermal infrared image,the pixel value of the pulse sensitive area of the human face is collected,and the human heart rate signal is detected to distinguish between real faces and spoofing faces.In order to better obtain the texture features of the face,an image fusion algorithm based on DTCWT and the improved Roberts algorithm is proposed.Firstly,DTCWT is used to decompose the thermal infrared image and visible light image of the face to obtain high-and low-frequency subbands.Then,the method based on region energy and the improved Roberts algorithm are then used to fuse the coefficients of the high-and low-frequency subbands.Finally,the DTCWT inverse transform is used to obtain the fused image containing the facial texture features.Face recognition is carried out on the fused image to realize identity authentication.Experimental results show that this algorithm can effectively resist attacks from photos,videos or masks.Compared with the use of visible light images alone for face recognition,this algorithm has higher recognition accuracy and better robustness.

Face Anti-Spoofing with Unknown Attacks:A Comprehensive Feature Extraction and Representation Perspective

Face anti-spoofing aims at detecting whether the input is a real photo of a user(living)or a fake(spoofing)image.As new types of attacks keep emerging,the detection of unknown attacks,known as Zero-Shot Face Anti-Spoofing(ZSFA),has become increasingly important in both academia and industry.Existing ZSFA methods mainly focus on extracting discriminative features between spoofing and living faces.However,the nature of the spoofing faces is to trick anti-spoofing systems by mimicking the livings,therefore the deceptive features between the known attacks and the livings,which have been ignored by existing ZSFA methods,are essential to comprehensively represent the livings.Therefore,existing ZSFA models are incapable of learning the complete representations of living faces and thus fall short of effectively detecting newly emerged attacks.To tackle this problem,we propose an innovative method that effectively captures both the deceptive and discriminative features distinguishing between genuine and spoofing faces.Our method consists of two main components:a two-against-all training strategy and a semantic autoencoder.The two-against-all training strategy is employed to separate deceptive and discriminative features.To address the subsequent invalidation issue of categorical functions and the dominance disequilibrium issue among different dimensions of features after importing deceptive features,we introduce a modified semantic autoencoder.This autoencoder is designed to map all extracted features to a semantic space,thereby achieving a balance in the dominance of each feature dimension.We combine our method with the feature extraction model ResNet50,and experimental results show that the trained ResNet50 model simultaneously achieves a feasible detection of unknown attacks and comparably accurate detection of known spoofing.Experimental results confirm the superiority and effectiveness of our proposed method in identifying the living with the interference of both known and unknown spoofing types.

Face anti-spoofing based on multi-modal and multi-scale features fusion

Face anti-spoofing is used to assist face recognition system to judge whether the detected face is real face or fake face. In the traditional face anti-spoofing methods, features extracted by hand are used to describe the difference between living face and fraudulent face. But these handmade features do not apply to different variations in an unconstrained environment. The convolutional neural network(CNN) for face deceptions achieves considerable results. However, most existing neural network-based methods simply use neural networks to extract single-scale features from single-modal data, while ignoring multi-scale and multi-modal information. To address this problem, a novel face anti-spoofing method based on multi-modal and multi-scale features fusion(MMFF) is proposed. Specifically, first residual network(Resnet)-34 is adopted to extract features of different scales from each modality, then these features of different scales are fused by feature pyramid network(FPN), finally squeeze-and-excitation fusion(SEF) module and self-attention network(SAN) are combined to fuse features from different modalities for classification. Experiments on the CASIA-SURF dataset show that the new method based on MMFF achieves better performance compared with most existing methods.

Anti-Spoofing:Integrated Information Authentication of BeiDou-ⅡCivil Navigation Message

The BeiDou-Ⅱcivil navigation message(BDⅡ-CNAV)is transmitted in an open environment and no information integrity protection measures are provided.Hence,the BDⅡ-CNAV faces the threat of spoofing attacks,which can lead to wrong location reports and time indication.In order to deal with this threat,we proposed a scheme of anti-spoofing for BDⅡ-CNAV based on integrated information authentication.This scheme generates two type authentication information,one is authentication code information(ACI),which is applied to confirm the authenticity and reliability of satellite time information,and the other is signature information,which is used to authenticate the integrity of satellite location information and other information.Both authentication information is designed to embed into the reserved bits in BDⅡ-CNAV without changing the frame structure.In order to avoid authentication failure caused by public key error or key error,the key or public key prompt information(KPKPI)are designed to remind the receiver to update both keys in time.Experimental results indicate that the scheme can successfully detect spoofing attacks,and the authentication delay is less than 1%of the transmission delay,which meets the requirements of BDⅡ-CNAV information authentication.

基于SSD与FaceNet的人脸识别系统设计

人脸识别技术广泛应用于考勤管理、移动支付等智慧建设中。伴随着常态化的口罩干扰,传统人脸识别算法已无法满足实际应用需求,为此,本文利用深度学习模型SSD以及FaceNet模型对人脸识别系统展开设计。首先,为消除现有数据集中亚洲人脸占比小造成的类内间距变化差距不明显的问题,在CAS-IA Web Face公开数据集的基础上对亚洲人脸数据进行扩充;其次,为解决不同口罩样式对特征提取的干扰,使用SSD人脸检测模型与DLIB人脸关键点检测模型提取人脸关键点,并利用人脸关键点与口罩的空间位置关系,额外随机生成不同的口罩人脸,组成混合数据集;最后,在混合数据集上进行模型训练并将训练好的模型移植到人脸识别系统中,进行检测速度与识别精度验证。实验结果表明,系统的实时识别速度达20 fps以上,人脸识别模型准确率在构建的混合数据集中达到97.1%,在随机抽取的部分LFW数据集验证的准确率达99.7%,故而该系统可满足实际应用需求,在一定程度上提高人脸识别的鲁棒性与准确性。

基于YOLO5Face重分布的小尺度人脸检测方法

针对复杂场景下小尺度人脸检测精度较低的问题,提出了一种基于YOLO5Face重分布的小尺度人脸检测方法。方法以YOLO5Face为基础,在网络浅层引入改进的CBAM注意力并对模型计算重分布,提升复杂场景下小尺度人脸检测精度的同时降低模型参数量;采用融合mixup的数据增强方法,充分训练模型小尺度人脸检测分支;依据人脸检测特性,将softmax损失作为分类损失以最大化类间特征的差异。在WiderFace各个子集上的实验结果表明,与主流人脸检测方法相比,改进后的模型满足实时性的同时,小尺度人脸检测精度较高,其中Hard子集检测精度比YOLO5Face提升2个百分点。

VR/AR-AdaptFace:面向虚拟现实与增强现实的自适应多模态面部替换模型

随着VR/AR技术的迅猛发展,用户对于沉浸式体验的需求日益增长。同时,虚拟人脸技术亦趋成熟。基于此,本文探索将高度拟真的虚拟人脸融入VR/AR,以增强用户体验的自然度与沉浸感。然而,在虚拟数字人领域,图像生成及换脸技术在VR/AR环境下仍遇诸多挑战,尤其是唇形合成模型在动态场景及多语言环境下的性能需进一步优化。为解决上述问题,本文提出VR/AR-AdaptFace模型,一个面向虚拟现实与增强现实的自适应多模态面部替换方案。该模型由两大模块构成:“文颜绘真”模块,采用先进的文本至图像转换技术和特定类别先验保存策略,优化虚拟人脸生成,并通过注意力机制大幅提升图像质量;“语唇映生”模块,依托强大的生成器、唇形同步判别器及视觉质量判别器,实现语音与唇形的精准同步,为VR/AR场景中的动态交互带来更加逼真的体验。

基于改进YOLOv5s-face的Face5系列人脸检测算法

针对人脸检测中小尺度人脸和遮挡人脸的漏检问题,提出了一种基于改进YOLOv5s-face(you only look once version 5 small-face)的Face5系列人脸检测算法Face5S(face5 small)和Face5M(face5 medium)。使用马赛克(mosaic)和图像混合(mixup)数据增强方法,提升算法在复杂场景下检测人脸的泛化性和稳定性;通过改进C3的网络结构和引入可变形卷积(DCNv2)降低算法的参数量,提高算法提取特征的灵活性;通过引入特征的内容感知重组上采样算子(CARAFE),提高多尺度人脸的检测性能;引入损失函数WIoUV3(wise intersection over union version 3),提升算法的小尺度人脸检测性能。实验结果表明,在WIDER FACE验证集上,相较于YOLOv5s-face算法,Face5S算法的平均mAP@0.5提升了1.03%;相较于先进的人脸检测算法ASFD-D3(automatic and scalable face detector-D3)和TinaFace,Face5M算法的平均mAP@0.5分别提升了1.07%和2.11%,提出的Face5系列算法能够有效提升算法对小尺度和部分遮挡人脸的检测性能,同时具有实时性。

Probabilistic analysis of tunnel face seismic stability in layered rock masses using Polynomial Chaos Kriging metamodel

Face stability is an essential issue in tunnel design and construction.Layered rock masses are typical and ubiquitous;uncertainties in rock properties always exist.In view of this,a comprehensive method,which combines the Upper bound Limit analysis of Tunnel face stability,the Polynomial Chaos Kriging,the Monte-Carlo Simulation and Analysis of Covariance method(ULT-PCK-MA),is proposed to investigate the seismic stability of tunnel faces.A two-dimensional analytical model of ULT is developed to evaluate the virtual support force based on the upper bound limit analysis.An efficient probabilistic analysis method PCK-MA based on the adaptive Polynomial Chaos Kriging metamodel is then implemented to investigate the parameter uncertainty effects.Ten input parameters,including geological strength indices,uniaxial compressive strengths and constants for three rock formations,and the horizontal seismic coefficients,are treated as random variables.The effects of these parameter uncertainties on the failure probability and sensitivity indices are discussed.In addition,the effects of weak layer position,the middle layer thickness and quality,the tunnel diameter,the parameters correlation,and the seismic loadings are investigated,respectively.The results show that the layer distributions significantly influence the tunnel face probabilistic stability,particularly when the weak rock is present in the bottom layer.The efficiency of the proposed ULT-PCK-MA is validated,which is expected to facilitate the engineering design and construction.

Rock mass quality prediction on tunnel faces with incomplete multi-source dataset via tree-augmented naive Bayesian network

Rock mass quality serves as a vital index for predicting the stability and safety status of rock tunnel faces.In tunneling practice,the rock mass quality is often assessed via a combination of qualitative and quantitative parameters.However,due to the harsh on-site construction conditions,it is rather difficult to obtain some of the evaluation parameters which are essential for the rock mass quality prediction.In this study,a novel improved Swin Transformer is proposed to detect,segment,and quantify rock mass characteristic parameters such as water leakage,fractures,weak interlayers.The site experiment results demonstrate that the improved Swin Transformer achieves optimal segmentation results and achieving accuracies of 92%,81%,and 86%for water leakage,fractures,and weak interlayers,respectively.A multisource rock tunnel face characteristic(RTFC)dataset includes 11 parameters for predicting rock mass quality is established.Considering the limitations in predictive performance of incomplete evaluation parameters exist in this dataset,a novel tree-augmented naive Bayesian network(BN)is proposed to address the challenge of the incomplete dataset and achieved a prediction accuracy of 88%.In comparison with other commonly used Machine Learning models the proposed BN-based approach proved an improved performance on predicting the rock mass quality with the incomplete dataset.By utilizing the established BN,a further sensitivity analysis is conducted to quantitatively evaluate the importance of the various parameters,results indicate that the rock strength and fractures parameter exert the most significant influence on rock mass quality.

广角SS-OCT En Face结构投射图揭示未成年人玻璃体早期液化特征

目的:应用广角扫频源光学相干断层扫描成像(swept-source optical coherence tomography,SS-OCT)的en face结构投射图研究玻璃体早期液化特征。方法:使用SS-OCT进行18 mm×18 mm的容积(Cube)扫描,创建并分析健康未成年人(年龄5~18岁)70眼的系列玻璃体en face结构投射图。结果:在未成年人中,视网膜前的玻璃体包含4种液化结构,分别为后皮质前玻璃体囊袋(posterior precortical vitreous pocket,PPVP)、视盘前Martegiani区(the area of Martegiani,AM)、血管前液化裂隙(prevascular vitreous fissures,PVF)和液化池(cistern)。所有研究眼均能检出PPVP、AM和PVF,其中22眼(31.4%)的PPVP和AM连通。41眼(58.6%)可检出液化池,且其年龄大于未检出液化池的个体(P=0.01),液化池的发生与年龄呈正相关(r_(s)=0.315,P=0.008)。液化池的象限空间分布频率依次为颞上(90.2%)、鼻上(58.5%)、颞下(36.6%)、鼻下(24.4%),最常累及颞上象限(P<0.001)。结论:PPVP、AM和PVF是健康人群视网膜前玻璃体早期液化过程中均出现的特征。液化池的发生与年龄呈正相关,最常出现在颞上象限,可能是年龄相关性玻璃体液化变性的结果。

VanillaFaceNet:一种高精度快速推理的牛脸识别方法

快速精准确定牛只身份对于牛只活体贷款,改善牛只骗保等问题具有重要意义。针对不同牛只面部差异小,FaceNet网络层数深,推理速度较慢,模型分类精度不足等问题,该研究提出了基于FaceNet的牛脸识别方法-VanillaFaceNet。该方法首先将FaceNet的主干特征提取网络替换为极简网络VanillaNet-13并提出动态激活和增强型线性变换的激活函数两种方法提高网络的非线性;然后,提出一种新的DBCA(dual-branch coordinate attention)注意力模块,能够更好地反映不同牛只面部特征之间的差异,从而提高网络的识别精度;最后,针对triplet loss仅能减小牛只类间差异的问题,采用center-triplet loss联合监督来减少牛只类内差异,从而提高了相同牛只身份比对的准确性。基于自建的牛脸数据集对该模型进行训练和测试,试验结果表明,VanillaFaceNet对牛只识别的准确率达到88.21%,每秒传输帧数为26.23帧。与FaceNet、MobileFaceNet、CenterFace、CosFace和ArcFace算法相比,本文算法的识别准确率分别提高了2.99、9.58、6.26、3.85和4.49个百分点,推理速度分别提升了2.67、0.77、0.10、1.28和0.94帧/s。该模型对牛只有较为优秀的识别效果,适于在嵌入式设备上部署,实现了牛只面部识别精度和推理速度之间的平衡。

Anti-Spoof Reliable Biometry of Fingerprints Using <i>En-Face</i>Optical Coherence Tomography

Optical coherence tomography (OCT) is a relatively new imaging technology which can produce high resolution images of three-dimensional structures. OCT has been mainly used for medical applications such as for ophthalmology and dermatology. In this study we demonstrate its capability in providing much more reliable biometry identification of fingerprints than conventional methods. We prove that OCT can serve secure control of genuine fingerprints as it can detect if extra layers are placed above the finger. This can prevent with a high probability, intruders to a secure area trying to foul standard systems based on imaging the finger surface. En-Face OCT method is employed and recommended for its capability of providing not only the axial succession of layers in depth, but the en-face image that allows the traditional pattern identification. Another reason for using such OCT technology is that it is compatible with dynamic focus and therefore can provide enhanced transversal resolution and sensitivity. Two En-Face OCT systems are used to evaluate the need for high resolution and conclusions are drawn in terms of the most potential commercial route to ex- ploitation.

Stability analysis of tunnel face reinforced with face bolts

Face bolting has been widely utilized to enhance the stability of tunnel face,particularly in soft soil tunnels.However,the influence of bolt reinforcement and its layout on tunnel face stability has not been systematically studied.Based on the theory of linear elastic mechanics,this study delved into the specific mechanisms of bolt reinforcement on the tunnel face in both horizontal and vertical dimensions.It also identified the primary failure types of bolts.Additionally,a design approach for tunnel face bolts that incorporates spatial layout was established using the limit equilibrium method to enhance the conventional wedge-prism model.The proposed model was subsequently validated through various means,and the specific influence of relevant bolt design parameters on tunnel face stability was analyzed.Furthermore,design principles for tunnel face bolts under different geological conditions were presented.The findings indicate that bolt failure can be categorized into three stages:tensile failure,pullout failure,and comprehensive failure.Increasing cohesion,internal friction angle,bolt density,and overlap length can effectively enhance tunnel face stability.Due to significant variations in stratum conditions,tailored design approaches based on specific failure stages are necessary for bolt design.

基于RetinaFace与FaceNet的动态人脸识别系统设计

针对在现有人脸静态识别过程中被识别人需等待配合的问题,文中提出了一种动态人脸识别系统。该系统采用了基于RetinaFace与FaceNet算法的动态人脸检测和识别方法,并进行了优化,以达到高识别精度和实时性的目标。其中,RetinaFace检测采用GhostNet作为骨干网络,使用Adaptive-NMS(Non Max Suppression)非极大值抑制用于人脸框的回归,FaceNet识别采用MobileNetV1作为骨干网络,使用Triplet损失与交叉熵损失结合的联合损失函数用以人脸分类。优化后的算法在检测与识别上具有良好表现,改进RetinaFace算法在WiderFace数据集下检测精度为93.35%、90.84%和80.43%,FPS(Frames Per Second)可达53 frame·s^(-1)。动态人脸检测平均检测精度为96%,FPS为21 frame·s^(-1)。当FaceNet阈值设为1.15时,识别率最高达到98.23%。动态识别系统平均识别精度98%,FPS可达20 frame·s^(-1)。实验结果表明,该系统解决了人脸静态识别中需等待配合的问题,具有较高的识别精度与实时性。

Generating animatable 3D cartoon faces from single portraits

Background With the development of virtual reality(VR)technology,there is a growing need for customized 3D avatars.However,traditional methods for 3D avatar modeling are either time-consuming or fail to retain the similarity to the person being modeled.This study presents a novel framework for generating animatable 3D cartoon faces from a single portrait image.Methods First,we transferred an input real-world portrait to a stylized cartoon image using StyleGAN.We then proposed a two-stage reconstruction method to recover a 3D cartoon face with detailed texture.Our two-stage strategy initially performs coarse estimation based on template models and subsequently refines the model by nonrigid deformation under landmark supervision.Finally,we proposed a semantic-preserving face-rigging method based on manually created templates and deformation transfer.Conclusions Compared with prior arts,the qualitative and quantitative results show that our method achieves better accuracy,aesthetics,and similarity criteria.Furthermore,we demonstrated the capability of the proposed 3D model for real-time facial animation.

Sparse representation scheme with enhanced medium pixel intensity for face recognition

Sparse representation is an effective data classification algorithm that depends on the known training samples to categorise the test sample.It has been widely used in various image classification tasks.Sparseness in sparse representation means that only a few of instances selected from all training samples can effectively convey the essential class-specific information of the test sample,which is very important for classification.For deformable images such as human faces,pixels at the same location of different images of the same subject usually have different intensities.Therefore,extracting features and correctly classifying such deformable objects is very hard.Moreover,the lighting,attitude and occlusion cause more difficulty.Considering the problems and challenges listed above,a novel image representation and classification algorithm is proposed.First,the authors’algorithm generates virtual samples by a non-linear variation method.This method can effectively extract the low-frequency information of space-domain features of the original image,which is very useful for representing deformable objects.The combination of the original and virtual samples is more beneficial to improve the clas-sification performance and robustness of the algorithm.Thereby,the authors’algorithm calculates the expression coefficients of the original and virtual samples separately using the sparse representation principle and obtains the final score by a designed efficient score fusion scheme.The weighting coefficients in the score fusion scheme are set entirely automatically.Finally,the algorithm classifies the samples based on the final scores.The experimental results show that our method performs better classification than conventional sparse representation algorithms.

The Relation between Mental Workload and Face Temperature in Flight Simulation

In this research, we study the relationship between mental workload and facial temperature of aircraft participants during a simulated takeoff flight. We conducted experiments to comprehend the correlation between work and facial temperature within the flight simulator. The experiment involved a group of 10 participants who played the role of pilots in a simulated A-320 flight. Six different flying scenarios were designed to simulate normal and emergency situations on airplane takeoff that would occur in different levels of mental workload for the participants. The measurements were workload assessment, face temperatures, and heart rate monitoring. Throughout the experiments, we collected a total of 120 instances of takeoffs, together with over 10 hours of time-series data including heart rate, workload, and face thermal images and temperatures. Comparative analysis of EEG data and thermal image types, revealed intriguing findings. The results indicate a notable inverse relationship between workload and facial muscle temperatures, as well as facial landmark points. The results of this study contribute to a deeper understanding of the physiological effects of workload, as well as practical implications for aviation safety and performance.