This research presents an improved real-time face recognition system at a low<span><span><span style="font-family:" color:red;"=""> </span></span></span><...This research presents an improved real-time face recognition system at a low<span><span><span style="font-family:" color:red;"=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">resolution of 15 pixels with pose and emotion and resolution variations. We have designed our datasets named LRD200 and LRD100, which have been used for training and classification. The face detection part uses the Viola-Jones algorithm, and the face recognition part receives the face image from the face detection part to process it using the Local Binary Pattern Histogram (LBPH) algorithm with preprocessing using contrast limited adaptive histogram equalization (CLAHE) and face alignment. The face database in this system can be updated via our custom-built standalone android app and automatic restarting of the training and recognition process with an updated database. Using our proposed algorithm, a real-time face recognition accuracy of 78.40% at 15</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px and 98.05% at 45</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px have been achieved using the LRD200 database containing 200 images per person. With 100 images per person in the database (LRD100) the achieved accuracies are 60.60% at 15</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px and 95% at 45</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px respectively. A facial deflection of about 30</span></span></span><span><span><span><span><span style="color:#4F4F4F;font-family:-apple-system, " font-size:16px;white-space:normal;background-color:#ffffff;"="">°</span></span><span> on either side from the front face showed an average face recognition precision of 72.25%-81.85%. This face recognition system can be employed for law enforcement purposes, where the surveillance camera captures a low-resolution image because of the distance of a person from the camera. It can also be used as a surveillance system in airports, bus stations, etc., to reduce the risk of possible criminal threats.</span></span></span></span>展开更多
With the rapid development of artificial intelligence,face recognition systems are widely used in daily lives.Face recognition applications often need to process large amounts of image data.Maintaining the accuracy an...With the rapid development of artificial intelligence,face recognition systems are widely used in daily lives.Face recognition applications often need to process large amounts of image data.Maintaining the accuracy and low latency is critical to face recognition systems.After analyzing the two-tier architecture“client-cloud”face recognition systems,it is found that these systems have high latency and network congestion when massive recognition requirements are needed to be responded,and it is very inconvenient and inefficient to deploy and manage relevant applications on the edge of the network.This paper proposes a flexible and efficient edge computing accelerated architecture.By offloading part of the computing tasks to the edge server closer to the data source,edge computing resources are used for image preprocessing to reduce the number of images to be transmitted,thus reducing the network transmission overhead.Moreover,the application code does not need to be rewritten and can be easily migrated to the edge server.We evaluate our schemes based on the open source Azure IoT Edge,and the experimental results show that the three-tier architecture“Client-Edge-Cloud”face recognition system outperforms the state-of-art face recognition systems in reducing the average response time.展开更多
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 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.展开更多
Corona virus(COVID-19)is once in a life time calamity that has resulted in thousands of deaths and security concerns.People are using face masks on a regular basis to protect themselves and to help reduce corona virus...Corona virus(COVID-19)is once in a life time calamity that has resulted in thousands of deaths and security concerns.People are using face masks on a regular basis to protect themselves and to help reduce corona virus transmission.During the on-going coronavirus outbreak,one of the major priorities for researchers is to discover effective solution.As important parts of the face are obscured,face identification and verification becomes exceedingly difficult.The suggested method is a transfer learning using MobileNet V2 based technology that uses deep feature such as feature extraction and deep learning model,to identify the problem of face masked identification.In the first stage,we are applying face mask detector to identify the face mask.Then,the proposed approach is applying to the datasets from Canadian Institute for Advanced Research10(CIFAR10),Modified National Institute of Standards and Technology Database(MNIST),Real World Masked Face Recognition Database(RMFRD),and Stimulated Masked Face Recognition Database(SMFRD).The proposed model is achieving recognition accuracy 99.82%with proposed dataset.This article employs the four pre-programmed models VGG16,VGG19,ResNet50 and ResNet101.To extract the deep features of faces with VGG16 is achieving 99.30%accuracy,VGG19 is achieving 99.54%accuracy,ResNet50 is achieving 78.70%accuracy and ResNet101 is achieving 98.64%accuracy with own dataset.The comparative analysis shows,that our proposed model performs better result in all four previous existing models.The fundamental contribution of this study is to monitor with face mask and without face mask to decreases the pace of corona virus and to detect persons using wearing face masks.展开更多
Face recognition has been a hot-topic in the field of pattern recognition where feature extraction and classification play an important role. However, convolutional neural network (CNN) and local binary pattern (LB...Face recognition has been a hot-topic in the field of pattern recognition where feature extraction and classification play an important role. However, convolutional neural network (CNN) and local binary pattern (LBP) can only extract single features of facial images, and fail to select the optimal classifier. To deal with the problem of classifier parameter optimization, two structures based on the support vector machine (SVM) optimized by artificial bee colony (ABC) algorithm are proposed to classify CNN and LBP features separately. In order to solve the single feature problem, a fusion system based on CNN and LBP features is proposed. The facial features can be better represented by extracting and fusing the global and local information of face images. We achieve the goal by fusing the outputs of feature classifiers. Explicit experimental results on Olivetti Research Laboratory (ORL) and face recognition technology (FERET) databases show the superiority of the proposed approaches.展开更多
Convolutional neural networks continually evolve to enhance accuracy in addressing various problems,leading to an increase in computational cost and model size.This paper introduces a novel approach for pruning face r...Convolutional neural networks continually evolve to enhance accuracy in addressing various problems,leading to an increase in computational cost and model size.This paper introduces a novel approach for pruning face recognition models based on convolutional neural networks.The proposed method identifies and removes inefficient filters based on the information volume in feature maps.In each layer,some feature maps lack useful information,and there exists a correlation between certain feature maps.Filters associated with these two types of feature maps impose additional computational costs on the model.By eliminating filters related to these categories of feature maps,the reduction of both computational cost and model size can be achieved.The approach employs a combination of correlation analysis and the summation of matrix elements within each feature map to detect and eliminate inefficient filters.The method was applied to two face recognition models utilizing the VGG16 and ResNet50V2 backbone architectures.In the proposed approach,the number of filters removed in each layer varies,and the removal process is independent of the adjacent layers.The convolutional layers of both backbone models were initialized with pre-trained weights from ImageNet.For training,the CASIA-WebFace dataset was utilized,and the Labeled Faces in the Wild(LFW)dataset was employed for benchmarking purposes.In the VGG16-based face recognition model,a 0.74%accuracy improvement was achieved while reducing the number of convolution parameters by 26.85%and decreasing Floating-point operations per second(FLOPs)by 47.96%.For the face recognition model based on the ResNet50V2 architecture,the ArcFace method was implemented.The removal of inactive filters in this model led to a slight decrease in accuracy by 0.11%.However,it resulted in enhanced training speed,a reduction of 59.38%in convolution parameters,and a 57.29%decrease in FLOPs.展开更多
Face recognition (FR) technology has numerous applications in artificial intelligence including biometrics, security,authentication, law enforcement, and surveillance. Deep learning (DL) models, notably convolutional ...Face recognition (FR) technology has numerous applications in artificial intelligence including biometrics, security,authentication, law enforcement, and surveillance. Deep learning (DL) models, notably convolutional neuralnetworks (CNNs), have shown promising results in the field of FR. However CNNs are easily fooled since theydo not encode position and orientation correlations between features. Hinton et al. envisioned Capsule Networksas a more robust design capable of retaining pose information and spatial correlations to recognize objects morelike the brain does. Lower-level capsules hold 8-dimensional vectors of attributes like position, hue, texture, andso on, which are routed to higher-level capsules via a new routing by agreement algorithm. This provides capsulenetworks with viewpoint invariance, which has previously evaded CNNs. This research presents a FR model basedon capsule networks that was tested using the LFW dataset, COMSATS face dataset, and own acquired photos usingcameras measuring 128 × 128 pixels, 40 × 40 pixels, and 30 × 30 pixels. The trained model outperforms state-ofthe-art algorithms, achieving 95.82% test accuracy and performing well on unseen faces that have been blurred orrotated. Additionally, the suggested model outperformed the recently released approaches on the COMSATS facedataset, achieving a high accuracy of 92.47%. Based on the results of this research as well as previous results, capsulenetworks perform better than deeper CNNs on unobserved altered data because of their special equivarianceproperties.展开更多
This paper uses principal component analysis (PCA) to train the face and extract the characteristic value. This approach achieves the purpose of rapid attendance. PCA is an early and important approach for face reco...This paper uses principal component analysis (PCA) to train the face and extract the characteristic value. This approach achieves the purpose of rapid attendance. PCA is an early and important approach for face recognization. It can reduce the dimension of face image space as well as describe the variation characteristics between different face images. The attendance system is a realtime system that requires shorter response time, for which PCA is a best choice. We use histogram equalization to eliminate the noise and improve the performance. With convenient MATLAB GUI visual operation interface, users can click on the corresponding button to implement face recognition tasks.展开更多
In many automatic face recognition systems, posture constraining is a key factor preventin g them from application. In thi5.paper, a series of strategles. will be described to achieve a system which enables face recog...In many automatic face recognition systems, posture constraining is a key factor preventin g them from application. In thi5.paper, a series of strategles. will be described to achieve a system which enables face recognition under varying pose. These approaches include the multi-view face modeling, the threshold image based face feature detection, the affine transformation based face posture normalization and the template matching based face idelltification. Combining all of these strategies, a face recognition system with the pose invariance is designed successfully. Using a 75MHZ Pentium PC and with a database of 75 individuals, 15 images for each person, and 225 test images with various postures, a very good recognition rate of 96.89% is obtained.展开更多
The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.H...The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.However,one’s negative psychology is often hidden,and it is difficult to detect and effectively intervene before the tragedy.How to effectively identify students with suicidal tendencies in order to prevent tragedies has aroused extensive research and discussion among the government,academia and the public.Through investigation and research,it is found that the current popular computer cutting-edge technologies such as artificial intelligence and computer vision can be well used for human emotion recognition and behavior prediction,and put into use in schools as a mental health auxiliary diagnosis and treatment system,thus effectively reducing the suicide rate on campus.The scenario assumes that machine learning can be used to deduce the risk of psychological problems in human samples by analyzing the frequency of negative emotions in facial expressions.Based on this,this paper proposes an effective solution for campus suicide prediction,and designs a set of auxiliary diagnosis and treatment system based on campus monitoring network system for suicide behavior prediction and student mental health analysis.Through preliminary experimental analysis and verification,the suicide psychological auxiliary diagnosis and treatment system has achieved good results in face recognition success rate,emotion recognition success rate and behavior prediction success rate.With the input of more experimental data and the increase of self-training time,the prediction system will perform better.展开更多
The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.H...The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.However,one's negative psychology is often hidden,and it is difficult to detect and effectively intervene before the tragedy.How to effectively identify students with suicidal tendencies in order to prevent tragedies has aroused extensive research and discussion among the government,academia and the public.Through investigation and research,it is found that the current popular computer cutting-edge technologies such as artificial intelligence and computer vision can be well used for human emotion recognition and behavior prediction,and put into use in schools as a mental health auxiliary diagnosis and treatment system,thus effectively reducing the suicide rate on campus.The scenario assumes that machine learning can be used to deduce the risk of psychological problems in human samples by analyzing the frequency of negative emotions in facial expressions.Based on this,this paper proposes an effective solution for campus suicide prediction,and designs a set of auxiliary diagnosis and treatment system based on campus monitoring network system for suicide behavior prediction and student mental health analysis.Through preliminary experimental analysis and verification,the suicide psychological auxiliary diagnosis and treatment system has achieved good results in face recognition success rate,emotion recognition success rate and behavior prediction success rate.With the input of more experimental data and the increase of self-training time,the prediction system will perform better.展开更多
Deep neural networks(DNNs)are widely adopted in daily life and the security problems of DNNs have drawn attention from both scientific researchers and industrial engineers.Many related works show that DNNs are vulnera...Deep neural networks(DNNs)are widely adopted in daily life and the security problems of DNNs have drawn attention from both scientific researchers and industrial engineers.Many related works show that DNNs are vulnerable to adversarial examples that are generated with subtle perturbation to original images in both digital domain and physical domain.As a most common application of DNNs,face recognition systems are likely to cause serious consequences if they are attacked by the adversarial examples.In this paper,we implement an adversarial attack system for face recognition in both digital domain that generates adversarial face images to fool the recognition system,and physical domain that generates customized glasses to fool the system when a person wears the glasses.Experiments show that our system attacks face recognition systems effectively.Furthermore,our system could misguide the recognition system to identify a person wearing the customized glasses as a certain target.We hope this research could help raise the attention of artificial intelligence security and promote building robust recognition systems.展开更多
Face recognition systems have enhanced human-computer interactions in the last ten years.However,the literature reveals that current techniques used for identifying or verifying faces are not immune to limitations.Pri...Face recognition systems have enhanced human-computer interactions in the last ten years.However,the literature reveals that current techniques used for identifying or verifying faces are not immune to limitations.Principal Component Analysis-Support Vector Machine(PCA-SVM)and Principal Component Analysis-Artificial Neural Network(PCA-ANN)are among the relatively recent and powerful face analysis techniques.Compared to PCA-ANN,PCA-SVM has demonstrated generalization capabilities in many tasks,including the ability to recognize objects with small or large data samples.Apart from requiring a minimal number of parameters in face detection,PCA-SVM minimizes generalization errors and avoids overfitting problems better than PCA-ANN.PCA-SVM,however,is ineffective and inefficient in detecting human faces in cases in which there is poor lighting,long hair,or items covering the subject’s face.This study proposes a novel PCASVM-based model to overcome the recognition problem of PCA-ANN and enhance face detection.The experimental results indicate that the proposed model provides a better face recognition outcome than PCA-SVM.展开更多
Support vector machine (SVM), as a novel approach in pattern recognition, has demonstrated a success in face detection and face recognition. In this paper, a face recognition approach based on the SVM classifier with ...Support vector machine (SVM), as a novel approach in pattern recognition, has demonstrated a success in face detection and face recognition. In this paper, a face recognition approach based on the SVM classifier with the nearest neighbor classifier (NNC) is proposed. The principal component analysis (PCA) is used to reduce the dimension and extract features. Then one-against-all stratedy is used to train the SVM classifiers. At the testing stage, we propose an al-展开更多
Face recognition provides a natural visual interface for human computer interaction (HCI) applications. The process of face recognition, however, is inhibited by variations in the appearance of face images caused by...Face recognition provides a natural visual interface for human computer interaction (HCI) applications. The process of face recognition, however, is inhibited by variations in the appearance of face images caused by changes in lighting, expression, viewpoint, aging and introduction of occlusion. Although various algorithms have been presented for face recognition, face recognition is still a very challenging topic. A novel approach of real time face recognition for HCI is proposed in the paper. In view of the limits of the popular approaches to foreground segmentation, wavelet multi-scale transform based background subtraction is developed to extract foreground objects. The optimal selection of the threshold is automatically determined, which does not require any complex supervised training or manual experimental calibration. A robust real time face recognition algorithm is presented, which combines the projection matrixes without iteration and kernel Fisher discriminant analysis (KFDA) to overcome some difficulties existing in the real face recognition. Superior performance of the proposed algorithm is demonstrated by comparing with other algorithms through experiments. The proposed algorithm can also be applied to the video image sequences of natural HCI.展开更多
Dimensionality reduction methods play an important role in face recognition. Principal component analysis(PCA) and two-dimensional principal component analysis(2DPCA) are two kinds of important methods in this field. ...Dimensionality reduction methods play an important role in face recognition. Principal component analysis(PCA) and two-dimensional principal component analysis(2DPCA) are two kinds of important methods in this field. Recent research seems like that 2DPCA method is superior to PCA method. To prove if this conclusion is always true, a comprehensive comparison study between PCA and 2DPCA methods was carried out. A novel concept, called column-image difference(CID), was proposed to analyze the difference between PCA and 2DPCA methods in theory. It is found that there exist some restrictive conditions when2 DPCA outperforms PCA. After theoretical analysis, the experiments were conducted on four famous face image databases. The experiment results confirm the validity of theoretical claim.展开更多
With the continuous progress of The Times and the development of technology,the rise of network social media has also brought the“explosive”growth of image data.As one of the main ways of People’s Daily communicati...With the continuous progress of The Times and the development of technology,the rise of network social media has also brought the“explosive”growth of image data.As one of the main ways of People’s Daily communication,image is widely used as a carrier of communication because of its rich content,intuitive and other advantages.Image recognition based on convolution neural network is the first application in the field of image recognition.A series of algorithm operations such as image eigenvalue extraction,recognition and convolution are used to identify and analyze different images.The rapid development of artificial intelligence makes machine learning more and more important in its research field.Use algorithms to learn each piece of data and predict the outcome.This has become an important key to open the door of artificial intelligence.In machine vision,image recognition is the foundation,but how to associate the low-level information in the image with the high-level image semantics becomes the key problem of image recognition.Predecessors have provided many model algorithms,which have laid a solid foundation for the development of artificial intelligence and image recognition.The multi-level information fusion model based on the VGG16 model is an improvement on the fully connected neural network.Different from full connection network,convolutional neural network does not use full connection method in each layer of neurons of neural network,but USES some nodes for connection.Although this method reduces the computation time,due to the fact that the convolutional neural network model will lose some useful feature information in the process of propagation and calculation,this paper improves the model to be a multi-level information fusion of the convolution calculation method,and further recovers the discarded feature information,so as to improve the recognition rate of the image.VGG divides the network into five groups(mimicking the five layers of AlexNet),yet it USES 3*3 filters and combines them as a convolution sequence.Network deeper DCNN,channel number is bigger.The recognition rate of the model was verified by 0RL Face Database,BioID Face Database and CASIA Face Image Database.展开更多
The purpose of this paper is to solve the problem of robust face recognition(FR) with single sample per person(SSPP). In the scenario of FR with SSPP, we present a novel model local robust sparse representation(LRSR) ...The purpose of this paper is to solve the problem of robust face recognition(FR) with single sample per person(SSPP). In the scenario of FR with SSPP, we present a novel model local robust sparse representation(LRSR) to tackle the problem of query images with various intra-class variations,e.g., expressions, illuminations, and occlusion. FR with SSPP is a very difficult challenge due to lacking of information to predict the possible intra-class variation of the query images.The key idea of the proposed method is to combine a local sparse representation model and a patch-based generic variation dictionary learning model to predict the possible facial intraclass variation of the query images. The experimental results on the AR database, Extended Yale B database, CMU-PIE database and LFW database show that the proposed method is robust to intra-class variations in FR with SSPP, and outperforms the state-of-art approaches.展开更多
Most face recognition techniques have been successful in dealing with high-resolution(HR) frontal face images. However, real-world face recognition systems are often confronted with the low-resolution(LR) face images ...Most face recognition techniques have been successful in dealing with high-resolution(HR) frontal face images. However, real-world face recognition systems are often confronted with the low-resolution(LR) face images with pose and illumination variations. This is a very challenging issue, especially under the constraint of using only a single gallery image per person.To address the problem, we propose a novel approach called coupled kernel-based enhanced discriminant analysis(CKEDA).CKEDA aims to simultaneously project the features from LR non-frontal probe images and HR frontal gallery ones into a common space where discrimination property is maximized.There are four advantages of the proposed approach: 1) by using the appropriate kernel function, the data becomes linearly separable, which is beneficial for recognition; 2) inspired by linear discriminant analysis(LDA), we integrate multiple discriminant factors into our objective function to enhance the discrimination property; 3) we use the gallery extended trick to improve the recognition performance for a single gallery image per person problem; 4) our approach can address the problem of matching LR non-frontal probe images with HR frontal gallery images,which is difficult for most existing face recognition techniques.Experimental evaluation on the multi-PIE dataset signifies highly competitive performance of our algorithm.展开更多
It has been demonstrated that the linear discriminant analysis (LDA) is an effective approach in face recognition tasks. However, due to the high dimensionality of an image space, many LDA based approaches first use t...It has been demonstrated that the linear discriminant analysis (LDA) is an effective approach in face recognition tasks. However, due to the high dimensionality of an image space, many LDA based approaches first use the principal component analysis (PCA) to project an image into a lower dimensional space, then perform the LDA transform to extract discriminant feature. But some useful discriminant information to the following LDA transform will be lost in the PCA step. To overcome these defects, a face recognition method based on the discrete cosine transform (DCT) and the LDA is proposed. First the DCT is used to achieve dimension reduction, then LDA transform is performed on the lower space to extract features. Two face databases are used to test our method and the correct recognition rates of 97.5% and 96.0% are obtained respectively. The performance of the proposed method is compared with that of the PCA+ LDA method and the results show that the method proposed outperforms the PCA+ LDA method.展开更多
文摘This research presents an improved real-time face recognition system at a low<span><span><span style="font-family:" color:red;"=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">resolution of 15 pixels with pose and emotion and resolution variations. We have designed our datasets named LRD200 and LRD100, which have been used for training and classification. The face detection part uses the Viola-Jones algorithm, and the face recognition part receives the face image from the face detection part to process it using the Local Binary Pattern Histogram (LBPH) algorithm with preprocessing using contrast limited adaptive histogram equalization (CLAHE) and face alignment. The face database in this system can be updated via our custom-built standalone android app and automatic restarting of the training and recognition process with an updated database. Using our proposed algorithm, a real-time face recognition accuracy of 78.40% at 15</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px and 98.05% at 45</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px have been achieved using the LRD200 database containing 200 images per person. With 100 images per person in the database (LRD100) the achieved accuracies are 60.60% at 15</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px and 95% at 45</span></span></span><span><span><span style="font-family:;" "=""> </span></span></span><span style="font-family:Verdana;"><span style="font-family:Verdana;"><span style="font-family:Verdana;">px respectively. A facial deflection of about 30</span></span></span><span><span><span><span><span style="color:#4F4F4F;font-family:-apple-system, " font-size:16px;white-space:normal;background-color:#ffffff;"="">°</span></span><span> on either side from the front face showed an average face recognition precision of 72.25%-81.85%. This face recognition system can be employed for law enforcement purposes, where the surveillance camera captures a low-resolution image because of the distance of a person from the camera. It can also be used as a surveillance system in airports, bus stations, etc., to reduce the risk of possible criminal threats.</span></span></span></span>
基金This work is supported by the National Key Research and Development Program of China under Grant(No.2016YFB1000302)the National Natural Science Foundation of China under Grant(No.61832020).
文摘With the rapid development of artificial intelligence,face recognition systems are widely used in daily lives.Face recognition applications often need to process large amounts of image data.Maintaining the accuracy and low latency is critical to face recognition systems.After analyzing the two-tier architecture“client-cloud”face recognition systems,it is found that these systems have high latency and network congestion when massive recognition requirements are needed to be responded,and it is very inconvenient and inefficient to deploy and manage relevant applications on the edge of the network.This paper proposes a flexible and efficient edge computing accelerated architecture.By offloading part of the computing tasks to the edge server closer to the data source,edge computing resources are used for image preprocessing to reduce the number of images to be transmitted,thus reducing the network transmission overhead.Moreover,the application code does not need to be rewritten and can be easily migrated to the edge server.We evaluate our schemes based on the open source Azure IoT Edge,and the experimental results show that the three-tier architecture“Client-Edge-Cloud”face recognition system outperforms the state-of-art face recognition systems in reducing the average response time.
文摘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.
文摘Corona virus(COVID-19)is once in a life time calamity that has resulted in thousands of deaths and security concerns.People are using face masks on a regular basis to protect themselves and to help reduce corona virus transmission.During the on-going coronavirus outbreak,one of the major priorities for researchers is to discover effective solution.As important parts of the face are obscured,face identification and verification becomes exceedingly difficult.The suggested method is a transfer learning using MobileNet V2 based technology that uses deep feature such as feature extraction and deep learning model,to identify the problem of face masked identification.In the first stage,we are applying face mask detector to identify the face mask.Then,the proposed approach is applying to the datasets from Canadian Institute for Advanced Research10(CIFAR10),Modified National Institute of Standards and Technology Database(MNIST),Real World Masked Face Recognition Database(RMFRD),and Stimulated Masked Face Recognition Database(SMFRD).The proposed model is achieving recognition accuracy 99.82%with proposed dataset.This article employs the four pre-programmed models VGG16,VGG19,ResNet50 and ResNet101.To extract the deep features of faces with VGG16 is achieving 99.30%accuracy,VGG19 is achieving 99.54%accuracy,ResNet50 is achieving 78.70%accuracy and ResNet101 is achieving 98.64%accuracy with own dataset.The comparative analysis shows,that our proposed model performs better result in all four previous existing models.The fundamental contribution of this study is to monitor with face mask and without face mask to decreases the pace of corona virus and to detect persons using wearing face masks.
基金supported by the Natural Science Foundation of Shandong Province ( ZR2014FM039)the National Natural Science Foundation of China ( 61771293)
文摘Face recognition has been a hot-topic in the field of pattern recognition where feature extraction and classification play an important role. However, convolutional neural network (CNN) and local binary pattern (LBP) can only extract single features of facial images, and fail to select the optimal classifier. To deal with the problem of classifier parameter optimization, two structures based on the support vector machine (SVM) optimized by artificial bee colony (ABC) algorithm are proposed to classify CNN and LBP features separately. In order to solve the single feature problem, a fusion system based on CNN and LBP features is proposed. The facial features can be better represented by extracting and fusing the global and local information of face images. We achieve the goal by fusing the outputs of feature classifiers. Explicit experimental results on Olivetti Research Laboratory (ORL) and face recognition technology (FERET) databases show the superiority of the proposed approaches.
文摘Convolutional neural networks continually evolve to enhance accuracy in addressing various problems,leading to an increase in computational cost and model size.This paper introduces a novel approach for pruning face recognition models based on convolutional neural networks.The proposed method identifies and removes inefficient filters based on the information volume in feature maps.In each layer,some feature maps lack useful information,and there exists a correlation between certain feature maps.Filters associated with these two types of feature maps impose additional computational costs on the model.By eliminating filters related to these categories of feature maps,the reduction of both computational cost and model size can be achieved.The approach employs a combination of correlation analysis and the summation of matrix elements within each feature map to detect and eliminate inefficient filters.The method was applied to two face recognition models utilizing the VGG16 and ResNet50V2 backbone architectures.In the proposed approach,the number of filters removed in each layer varies,and the removal process is independent of the adjacent layers.The convolutional layers of both backbone models were initialized with pre-trained weights from ImageNet.For training,the CASIA-WebFace dataset was utilized,and the Labeled Faces in the Wild(LFW)dataset was employed for benchmarking purposes.In the VGG16-based face recognition model,a 0.74%accuracy improvement was achieved while reducing the number of convolution parameters by 26.85%and decreasing Floating-point operations per second(FLOPs)by 47.96%.For the face recognition model based on the ResNet50V2 architecture,the ArcFace method was implemented.The removal of inactive filters in this model led to a slight decrease in accuracy by 0.11%.However,it resulted in enhanced training speed,a reduction of 59.38%in convolution parameters,and a 57.29%decrease in FLOPs.
基金Princess Nourah bint Abdulrahman University Riyadh,Saudi Arabia with Researchers Supporting Project Number:PNURSP2024R234.
文摘Face recognition (FR) technology has numerous applications in artificial intelligence including biometrics, security,authentication, law enforcement, and surveillance. Deep learning (DL) models, notably convolutional neuralnetworks (CNNs), have shown promising results in the field of FR. However CNNs are easily fooled since theydo not encode position and orientation correlations between features. Hinton et al. envisioned Capsule Networksas a more robust design capable of retaining pose information and spatial correlations to recognize objects morelike the brain does. Lower-level capsules hold 8-dimensional vectors of attributes like position, hue, texture, andso on, which are routed to higher-level capsules via a new routing by agreement algorithm. This provides capsulenetworks with viewpoint invariance, which has previously evaded CNNs. This research presents a FR model basedon capsule networks that was tested using the LFW dataset, COMSATS face dataset, and own acquired photos usingcameras measuring 128 × 128 pixels, 40 × 40 pixels, and 30 × 30 pixels. The trained model outperforms state-ofthe-art algorithms, achieving 95.82% test accuracy and performing well on unseen faces that have been blurred orrotated. Additionally, the suggested model outperformed the recently released approaches on the COMSATS facedataset, achieving a high accuracy of 92.47%. Based on the results of this research as well as previous results, capsulenetworks perform better than deeper CNNs on unobserved altered data because of their special equivarianceproperties.
基金Supported by Higher School Science and Technology Innovation Fund Project(2013160)Changzhi College Teaching Reform Fund Project(JY201503)
文摘This paper uses principal component analysis (PCA) to train the face and extract the characteristic value. This approach achieves the purpose of rapid attendance. PCA is an early and important approach for face recognization. It can reduce the dimension of face image space as well as describe the variation characteristics between different face images. The attendance system is a realtime system that requires shorter response time, for which PCA is a best choice. We use histogram equalization to eliminate the noise and improve the performance. With convenient MATLAB GUI visual operation interface, users can click on the corresponding button to implement face recognition tasks.
文摘In many automatic face recognition systems, posture constraining is a key factor preventin g them from application. In thi5.paper, a series of strategles. will be described to achieve a system which enables face recognition under varying pose. These approaches include the multi-view face modeling, the threshold image based face feature detection, the affine transformation based face posture normalization and the template matching based face idelltification. Combining all of these strategies, a face recognition system with the pose invariance is designed successfully. Using a 75MHZ Pentium PC and with a database of 75 individuals, 15 images for each person, and 225 test images with various postures, a very good recognition rate of 96.89% is obtained.
文摘The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.However,one’s negative psychology is often hidden,and it is difficult to detect and effectively intervene before the tragedy.How to effectively identify students with suicidal tendencies in order to prevent tragedies has aroused extensive research and discussion among the government,academia and the public.Through investigation and research,it is found that the current popular computer cutting-edge technologies such as artificial intelligence and computer vision can be well used for human emotion recognition and behavior prediction,and put into use in schools as a mental health auxiliary diagnosis and treatment system,thus effectively reducing the suicide rate on campus.The scenario assumes that machine learning can be used to deduce the risk of psychological problems in human samples by analyzing the frequency of negative emotions in facial expressions.Based on this,this paper proposes an effective solution for campus suicide prediction,and designs a set of auxiliary diagnosis and treatment system based on campus monitoring network system for suicide behavior prediction and student mental health analysis.Through preliminary experimental analysis and verification,the suicide psychological auxiliary diagnosis and treatment system has achieved good results in face recognition success rate,emotion recognition success rate and behavior prediction success rate.With the input of more experimental data and the increase of self-training time,the prediction system will perform better.
文摘The phenomenon of teenage campus suicide has become the focus of attention of parents,schools and the society.The causes behind it are extremely complicated,and the root cause is psychological and spiritual problems.However,one's negative psychology is often hidden,and it is difficult to detect and effectively intervene before the tragedy.How to effectively identify students with suicidal tendencies in order to prevent tragedies has aroused extensive research and discussion among the government,academia and the public.Through investigation and research,it is found that the current popular computer cutting-edge technologies such as artificial intelligence and computer vision can be well used for human emotion recognition and behavior prediction,and put into use in schools as a mental health auxiliary diagnosis and treatment system,thus effectively reducing the suicide rate on campus.The scenario assumes that machine learning can be used to deduce the risk of psychological problems in human samples by analyzing the frequency of negative emotions in facial expressions.Based on this,this paper proposes an effective solution for campus suicide prediction,and designs a set of auxiliary diagnosis and treatment system based on campus monitoring network system for suicide behavior prediction and student mental health analysis.Through preliminary experimental analysis and verification,the suicide psychological auxiliary diagnosis and treatment system has achieved good results in face recognition success rate,emotion recognition success rate and behavior prediction success rate.With the input of more experimental data and the increase of self-training time,the prediction system will perform better.
基金This work is supported in part by the National Natural Science Foundation of China under Grant 61902082,U1636215the Guangdong Province Key research and Development Plan under Grant 2019B010136003.
文摘Deep neural networks(DNNs)are widely adopted in daily life and the security problems of DNNs have drawn attention from both scientific researchers and industrial engineers.Many related works show that DNNs are vulnerable to adversarial examples that are generated with subtle perturbation to original images in both digital domain and physical domain.As a most common application of DNNs,face recognition systems are likely to cause serious consequences if they are attacked by the adversarial examples.In this paper,we implement an adversarial attack system for face recognition in both digital domain that generates adversarial face images to fool the recognition system,and physical domain that generates customized glasses to fool the system when a person wears the glasses.Experiments show that our system attacks face recognition systems effectively.Furthermore,our system could misguide the recognition system to identify a person wearing the customized glasses as a certain target.We hope this research could help raise the attention of artificial intelligence security and promote building robust recognition systems.
文摘Face recognition systems have enhanced human-computer interactions in the last ten years.However,the literature reveals that current techniques used for identifying or verifying faces are not immune to limitations.Principal Component Analysis-Support Vector Machine(PCA-SVM)and Principal Component Analysis-Artificial Neural Network(PCA-ANN)are among the relatively recent and powerful face analysis techniques.Compared to PCA-ANN,PCA-SVM has demonstrated generalization capabilities in many tasks,including the ability to recognize objects with small or large data samples.Apart from requiring a minimal number of parameters in face detection,PCA-SVM minimizes generalization errors and avoids overfitting problems better than PCA-ANN.PCA-SVM,however,is ineffective and inefficient in detecting human faces in cases in which there is poor lighting,long hair,or items covering the subject’s face.This study proposes a novel PCASVM-based model to overcome the recognition problem of PCA-ANN and enhance face detection.The experimental results indicate that the proposed model provides a better face recognition outcome than PCA-SVM.
基金This project was supported by Shanghai Shu Guang Project.
文摘Support vector machine (SVM), as a novel approach in pattern recognition, has demonstrated a success in face detection and face recognition. In this paper, a face recognition approach based on the SVM classifier with the nearest neighbor classifier (NNC) is proposed. The principal component analysis (PCA) is used to reduce the dimension and extract features. Then one-against-all stratedy is used to train the SVM classifiers. At the testing stage, we propose an al-
基金supported by the National Natural Science Foundation of China (Grant No.60872117)the Leading Academic Discipline Project of Shanghai Municipal Education Commission (Grant No.J50104)
文摘Face recognition provides a natural visual interface for human computer interaction (HCI) applications. The process of face recognition, however, is inhibited by variations in the appearance of face images caused by changes in lighting, expression, viewpoint, aging and introduction of occlusion. Although various algorithms have been presented for face recognition, face recognition is still a very challenging topic. A novel approach of real time face recognition for HCI is proposed in the paper. In view of the limits of the popular approaches to foreground segmentation, wavelet multi-scale transform based background subtraction is developed to extract foreground objects. The optimal selection of the threshold is automatically determined, which does not require any complex supervised training or manual experimental calibration. A robust real time face recognition algorithm is presented, which combines the projection matrixes without iteration and kernel Fisher discriminant analysis (KFDA) to overcome some difficulties existing in the real face recognition. Superior performance of the proposed algorithm is demonstrated by comparing with other algorithms through experiments. The proposed algorithm can also be applied to the video image sequences of natural HCI.
基金Projects(50275150,61173052)supported by the National Natural Science Foundation of China
文摘Dimensionality reduction methods play an important role in face recognition. Principal component analysis(PCA) and two-dimensional principal component analysis(2DPCA) are two kinds of important methods in this field. Recent research seems like that 2DPCA method is superior to PCA method. To prove if this conclusion is always true, a comprehensive comparison study between PCA and 2DPCA methods was carried out. A novel concept, called column-image difference(CID), was proposed to analyze the difference between PCA and 2DPCA methods in theory. It is found that there exist some restrictive conditions when2 DPCA outperforms PCA. After theoretical analysis, the experiments were conducted on four famous face image databases. The experiment results confirm the validity of theoretical claim.
文摘With the continuous progress of The Times and the development of technology,the rise of network social media has also brought the“explosive”growth of image data.As one of the main ways of People’s Daily communication,image is widely used as a carrier of communication because of its rich content,intuitive and other advantages.Image recognition based on convolution neural network is the first application in the field of image recognition.A series of algorithm operations such as image eigenvalue extraction,recognition and convolution are used to identify and analyze different images.The rapid development of artificial intelligence makes machine learning more and more important in its research field.Use algorithms to learn each piece of data and predict the outcome.This has become an important key to open the door of artificial intelligence.In machine vision,image recognition is the foundation,but how to associate the low-level information in the image with the high-level image semantics becomes the key problem of image recognition.Predecessors have provided many model algorithms,which have laid a solid foundation for the development of artificial intelligence and image recognition.The multi-level information fusion model based on the VGG16 model is an improvement on the fully connected neural network.Different from full connection network,convolutional neural network does not use full connection method in each layer of neurons of neural network,but USES some nodes for connection.Although this method reduces the computation time,due to the fact that the convolutional neural network model will lose some useful feature information in the process of propagation and calculation,this paper improves the model to be a multi-level information fusion of the convolution calculation method,and further recovers the discarded feature information,so as to improve the recognition rate of the image.VGG divides the network into five groups(mimicking the five layers of AlexNet),yet it USES 3*3 filters and combines them as a convolution sequence.Network deeper DCNN,channel number is bigger.The recognition rate of the model was verified by 0RL Face Database,BioID Face Database and CASIA Face Image Database.
基金supported in part by the National Natural Science Foundation of China(61673402,61273270,60802069)the Natural Science Foundation of Guangdong Province(2017A030311029,2016B010109002,2015B090912001,2016B010123005,2017B090909005)+1 种基金the Science and Technology Program of Guangzhou of China(201704020180,201604020024)the Fundamental Research Funds for the Central Universities of China
文摘The purpose of this paper is to solve the problem of robust face recognition(FR) with single sample per person(SSPP). In the scenario of FR with SSPP, we present a novel model local robust sparse representation(LRSR) to tackle the problem of query images with various intra-class variations,e.g., expressions, illuminations, and occlusion. FR with SSPP is a very difficult challenge due to lacking of information to predict the possible intra-class variation of the query images.The key idea of the proposed method is to combine a local sparse representation model and a patch-based generic variation dictionary learning model to predict the possible facial intraclass variation of the query images. The experimental results on the AR database, Extended Yale B database, CMU-PIE database and LFW database show that the proposed method is robust to intra-class variations in FR with SSPP, and outperforms the state-of-art approaches.
基金supported by National Natural Science Foundation of China(60802069,61273270)the Fundamental Research Funds for the Central Universities of China+1 种基金Natural Science Foundation of Guangdong Province(2014A030313173)Science and Technology Program of Guangzhou(2014Y2-00165,2014J4100114,2014J4100095)
文摘Most face recognition techniques have been successful in dealing with high-resolution(HR) frontal face images. However, real-world face recognition systems are often confronted with the low-resolution(LR) face images with pose and illumination variations. This is a very challenging issue, especially under the constraint of using only a single gallery image per person.To address the problem, we propose a novel approach called coupled kernel-based enhanced discriminant analysis(CKEDA).CKEDA aims to simultaneously project the features from LR non-frontal probe images and HR frontal gallery ones into a common space where discrimination property is maximized.There are four advantages of the proposed approach: 1) by using the appropriate kernel function, the data becomes linearly separable, which is beneficial for recognition; 2) inspired by linear discriminant analysis(LDA), we integrate multiple discriminant factors into our objective function to enhance the discrimination property; 3) we use the gallery extended trick to improve the recognition performance for a single gallery image per person problem; 4) our approach can address the problem of matching LR non-frontal probe images with HR frontal gallery images,which is difficult for most existing face recognition techniques.Experimental evaluation on the multi-PIE dataset signifies highly competitive performance of our algorithm.
文摘It has been demonstrated that the linear discriminant analysis (LDA) is an effective approach in face recognition tasks. However, due to the high dimensionality of an image space, many LDA based approaches first use the principal component analysis (PCA) to project an image into a lower dimensional space, then perform the LDA transform to extract discriminant feature. But some useful discriminant information to the following LDA transform will be lost in the PCA step. To overcome these defects, a face recognition method based on the discrete cosine transform (DCT) and the LDA is proposed. First the DCT is used to achieve dimension reduction, then LDA transform is performed on the lower space to extract features. Two face databases are used to test our method and the correct recognition rates of 97.5% and 96.0% are obtained respectively. The performance of the proposed method is compared with that of the PCA+ LDA method and the results show that the method proposed outperforms the PCA+ LDA method.