Secure Face Recognition Research Articles

Domain Generalization via Ensemble Stacking for Face Presentation Attack Detection

AbstractFace presentation attack detection (PAD) plays a pivotal role in securing face recognition systems against spoofing attacks. Although great progress has been made in designing face PAD methods, developing a model that can generalize well to unseen test domains remains a significant challenge. Moreover, due to the different types of spoofing attacks, creating a dataset with a sufficient number of samples for training deep neural networks is a laborious task. This work proposes a comprehensive solution that combines synthetic data generation and deep ensemble learning to enhance the generalization capabilities of face PAD. Specifically, synthetic data is generated by blending a static image with spatiotemporal-encoded images using alpha composition and video distillation. In this way, we simulate motion blur with varying alpha values, thereby generating diverse subsets of synthetic data that contribute to a more enriched training set. Furthermore, multiple base models are trained on each subset of synthetic data using stacked ensemble learning. This allows the models to learn complementary features and representations from different synthetic subsets. The meta-features generated by the base models are used as input for a new model called the meta-model. The latter combines the predictions from the base models, leveraging their complementary information to better handle unseen target domains and enhance overall performance. Experimental results from seven datasets—WMCA, CASIA-SURF, OULU-NPU, CASIA-MFSD, Replay-Attack, MSU-MFSD, and SiW-Mv2—highlight the potential to enhance presentation attack detection by using large-scale synthetic data and a stacking-based ensemble approach.

Artificial Immune System of Secure Face Recognition Against Adversarial Attacks

Artificial Immune System of Secure Face Recognition Against Adversarial Attacks

Integrating IoT and Python for Enhanced Face Recognition Security Systems

Integrating IoT and Python for Enhanced Face Recognition Security Systems

E-passport security systems and attack implications

Recent technological advances aim to recognise user data using biometrics such as the face, fingerprint, hand veins and iris. Currently, face prints are widely used to verify user data in e-passports. As a result, institutions face substantial difficulties in maintaining an appropriate level of security. Human error can introduce flaws that undermine security mechanisms. One potential solution to this problem is to install a facial recognition security system. Both hardware and software components make up this system, with the hardware being a camera and the software comprising face detection and identification algorithms.
 
 The purpose of this essay is to provide a thorough understanding of the Face Recognition Security System, including its application and deployment. Furthermore, the essay investigates the various weaknesses and methods of attack that could be used to target the system. The purpose of addressing these factors is to improve the effectiveness and robustness of system security, notably e-passport security.

Face recognition system with hybrid template protection scheme for Cyber–Physical-Social Services

This paper presents a secure face recognition system with advanced template protection schemes for Cyber–Physical–Social Services (CPSS). The implementation of the proposed system consists of five components. The initial step performs image preprocessing, where it detects the facial region from the captured image using the Tree-Structured Part Model (TSPM). The second phase involves feature extraction, where it utilizes the Scale Invariant Feature Transform (SIFT) descriptor to extract features from small patches of the preprocessed images, forming a collection of feature descriptors. The collection of feature descriptors is then clustered using the K-means clustering algorithm, returning the centers of K-clusters that serve as the vocabulary of a dictionary. Finally, a histogram is generated using the vocabularies and frequencies, referred to as the “Bag of Visual Words (BoVW)”. Using this dictionary and a feature learning technique called Sparse Representation Coding (SRC), followed by Spatial Pyramid Mapping (SPM), the system generates feature vectors from training/testing image samples. In the third component, the modified FaceHashing technique is applied to the original feature vectors, generating cancelable feature vectors. The fourth component employs a Bio-Cryptographic technique to preserve the cancelable feature vectors in a database. Lastly, the fifth component utilizes a multi-class linear SVM classifier on the decrypted and query-cancellable feature vector to classify users. The system evaluates its performance using FERET and CASIA-FaceV5 benchmark databases, providing 100% identification accuracy for 200-dimensional cancelable feature vectors. The performance and security comparisons demonstrate the superiority of the proposed system over existing methods.

Improving Face Anti-spoofing via Advanced Multi-perspective Feature Learning

Face anti-spoofing (FAS) plays a vital role in securing face recognition systems. Previous approaches usually learn spoofing features from a single perspective, in which only universal cues shared by all attack types are explored. However, such single-perspective-based approaches ignore the differences among various attacks and commonness between certain attacks and bona fides, thus tending to neglect some non-universal cues that contain strong discernibility against certain types. As a result, when dealing with multiple types of attacks, the above approaches may suffer from the uncomprehensive representation of bona fides and spoof faces. In this work, we propose a novel Advanced Multi-Perspective Feature Learning network (AMPFL), in which multiple perspectives are adopted to learn discriminative features, to improve the performance of FAS. Specifically, the proposed network first learns universal cues and several perspective-specific cues from multiple perspectives, then aggregates the above features and further enhances them to perform face anti-spoofing. In this way, AMPFL obtains features that are difficult to be captured by single-perspective-based methods and provides more comprehensive information on bona fides and spoof faces, thus achieving better performance for FAS. Experimental results show that our AMPFL achieves promising results in public databases, and it effectively solves the issues of single-perspective-based approaches.

Enhancement in ATM Machine Facility Using Face Recognition Security and OTP with Shuffle Keyboard

Abstract: To enhance the security of the ATM. To avoid ATM robberies and the wrong person misusing the ATM so that they can lead a secure and reliable life. The system is to support the intelligence system. Without hesitation, use ATMs to make the world a conversion community. Once the client inserts the cardboard into the ATM, a session is processed, so the system can start a face detection victimization camera placed close to the ATM and builds short-lived identity information of the clients, and the user face verification to perform on the ATM. A good user would use the traditional method because the invalid user cannot access the ATM card, they offer the secondary identification to the system mechanically, and the unauthorized person would continue the transactions. The objective of the proposed study in this paper is to prevent ATM fraud as well as secure transactions from the user side.

Brightness Based Password Safeguard System with Face Recognition in ATM

the importance of security in the authentication process as well as the increase in threat level posed by such malware has attracted many researchers to the field. Many attacks are successful in accessing social network accounts since the current password-based authentication paradigms are not efficient and robust enough as well as vulnerable to automated attacks. The traditional two-factor authentication mechanisms are not applicable to online social networks because physical token or biometric data cannot be easily used to log into users’ profiles. The simplest alternative is complementing the single factor (password-based) authentication process with additional identification elements, such as one-time PIN codes, generated by the user’s own device (e.g. the smart phone) or received via SMS. Proposed a brightness based authentication mechanism (i.e., Bright Pass) capable of enhancing the security of identity confirmation PIN codes without asking the user to memorize an additional secret value or to solve a complex cognitive task. This method introduces a new input value that is changed at every usage combining a something you know element (i.e., the PIN) with an interface element that cannot be captured by spyware, i.e., a bright or dark circle displayed on the phone screen to tell the user when to digit the correct PIN digit and when to digit a fake one. It prevents the malware from correctly inserting the PIN code, thereby disallowing the possibility to perform critical operations without the user’s agreement. Proposed work also focuses on implementing secure face recognition approach for user authentication. This approach will enhance the performance of ATM system

A Secure Face Recognition for IoT-enabled Healthcare System

In Healthcare, the Internet of Things (IoT)-enabled surveillance cameras capture thousands of images every day, where face recognition provides reliable security as well as smart treatment through patient sentiment analysis, emotion detection, automated nurse calls, and hospital traffic systems. In this article, a secure face recognition system for the IoT-enabled Healthcare system has been proposed. Here each registered person will be identified by his/her face biometric with strong template protection schemes. To protect the biometric information, three-step template protection techniques are proposed: (i) Cancelable biometrics , (ii) BioCrypto-Circuit , and (iii) BioCrypto-Protection . The performance of the proposed system has been tested on four benchmark face databases, CVL, IITK, Casia-Face-v5, and FERET. The results of the proposed system are reported in terms of the correct recognition rate and the equal error rate. These performances have also been compared with some state-of-the-art methods with respect to each employed database, which shows the novelty of the proposed system.

Smart Recognition System for High Security Zone

Abstract: In today’s world almost all the big institutions face significant security issues; thus, they need a few uniquely prepared staff to satisfy the necessary security. These work force, as being human, commit errors which might influence the degree of safety. A proposed answer for the previously mentioned problem could likewise be a Face Recognition Security System, which can distinguish gate crashers to confined or high security regions and help in limiting human mistake. So far, we have studied the image comparison techniques and tested the same against our dataset of sample images. Our program can almost accurately separate two similar or dissimilar images. We have further tested our sample images against different backgrounds, angles, size etc. and observed the output. We have also done the facial recognition part, i.e., we are able to detect faces and match it with that face from our database and show the result. Finally, we have done iris recognition by comparing static images of iris which gives us an output by showing whether the iris matches or not. Also, in this report we have discussed about our future plan.

EFFECTIVE APPROACH FOR FACE RECOGNITION AND ACTIVE SHAPE 3D MODELS USING KERNEL PRINCIPAL COMPONENT ANALYSIS

Face Recognition is a computer application that is capable of detecting, tracking, identifying or verifying human faces from an image or video captured using a digital camera. Although lot of progress has been made in domain of face detection and recognition for security, identification and surveillance purpose, but still there are issues hindering the progress to reach or surpass human level accuracy. These issues are variations in human facial appearance such as; varying lighting condition, noise in face images, scale, pose etc. Kernel principal component analysis (KPCA) as a powerful nonlinear feature extraction method has proven as a preprocessing step for classification algorithm. A face recognition approach based on KPCA and genetic algorithms (GAs) is proposed. By the use of the polynomial functions as a kernel function in KPCA, the high order relationships can be utilized and the nonlinear principal components can be obtained. After that nonlinear principal components, we use GAs to select the optimal feature set for classification.

Spoofing Face Detection Using Novel Edge-Net Autoencoder for Security

Recent security applications in mobile technologies and computer systems use face recognition for high-end security. Despite numerous security techniques, face recognition is considered a high-security control. Developers fuse and carry out face identification as an access authority into these applications. Still, face identification authentication is sensitive to attacks with a 2-D photo image or captured video to access the system as an authorized user. In the existing spoofing detection algorithm, there was some loss in the recreation of images. This research proposes an unobtrusive technique to detect face spoofing attacks that apply a single frame of the sequenced set of frames to overcome the above-said problems. This research offers a novel Edge-Net autoencoder to select convoluted and dominant features of the input diffused structure. First, this proposed method is tested with the Cross-ethnicity Face Anti-spoofing (CASIA), Fetal alcohol spectrum disorders (FASD) dataset. This database has three models of attacks: distorted photographs in printed form, photographs with removed eyes portion, and video attacks. The images are taken with three different quality cameras: low, average, and high-quality real and spoofed images. An extensive experimental study was performed with CASIA-FASD, 3 Diagnostic Machine Aid-Digital (DMAD) dataset that proved higher results when compared to existing algorithms.

Consistency Regularization for Deep Face Anti-Spoofing

Face anti-spoofing (FAS) plays a crucial role in securing face recognition systems. Empirically, given an image, a model with more consistent output on different views (i.e., augmentations) of this image usually performs better. Motivated by this exciting observation, we conjecture that encouraging feature consistency of different views may be a promising way to boost FAS models. In this paper, we explore this way thoroughly by enhancing both Embedding-level and Prediction-level Consistency Regularization (EPCR) in FAS. Specifically, at the embedding level, we design a dense similarity loss to maximize the similarities between all positions of two intermediate feature maps in a self-supervised fashion; while at the prediction level, we optimize the mean square error between the predictions of two views. Notably, our EPCR is free of annotations and can directly integrate into semi-supervised learning schemes. Considering different application scenarios, we further design five diverse semi-supervised protocols to measure semi-supervised FAS techniques. We conduct extensive experiments to show that EPCR can significantly improve the performance of several supervised and semi-supervised tasks on benchmark datasets. The codes and protocols are available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/clks-wzz/EPCR</uri> .

Application of Three Convolutional Neural Network Algorithms for Occluded Face Identification and Recognition for System Security

Deep Learning techniques in computer vision have become indispensable elements in biometric systems, especially face recognition. Facial recognition can be reliably used as an identification and authentication tool for premises or network access security. The masks wearing, which is one of the problems of concealment, are nowadays part of our habits for preventing COVID-19 disease, which leads to an obstruction of facial recognition. Occulted face recognition is one of the most challenging problems biometrics deals with. This paper presents convolution neural network algorithms for occluded face recognition. Our study presents a robust method using algorithms such as ResNet-50, VGG-19, and DenseNet-201 to contribute to occluded face recognition. Various parameters are used for this experiment, such as the cross-entropy used as a loss function and optimization algorithms adapted to deep learning. These include the SGD, Adam, and RMSProp optimizers. The convolution neural network algorithms were evaluated on the AR database. This experiment gave results that ranged from 94.81 to 99.81% for SGD, from 0 to 96.92 for Adam, and finally from 0 to 96.92 for RMSProp. DenseNet-201 algorithm using the SGD optimizer obtained the best score with 99.81%, and all the performance metrics used such as accuracy, MSE, F-score, recall, and MCC were used to confirm this good performance.

Implementation of face recognition system using BioCryptosystem as template protection scheme

This paper has designed and implemented a secure face recognition system using novel Bio-Cryptographic template protection schemes. The implementation of the proposed system has been divided into three components. The face regions are detected from the input images in the first component. Then some discriminant and distinctive feature extraction techniques are applied to extract features from the facial region. In the second component, the extracted features undergo a classification task to verify or identify the person based on their face biometric. The Huffman coding technique has been incorporated with the biometric face feature as BioCryptosystem for template protection scheme in the third component. The proposed BioCryptosystem works at the feature level in two steps: dictionary-based and user-key-based approaches for encryption followed by decryption processes. During experimentation, three benchmark facial databases, CVL, CASIA-FaceV5, and FERET have been employed. The performances have been compared with some existing state-of-the-art methods concerning each database which shows the superiority of the proposed system. Moreover, some comparisons have been performed for the security analysis of the employed BioCryptosystem, which shows that the proposed system takes less time for authentication and generates longer and stronger keys than other existing Bio-cryptographic techniques.

Liveness Detection Based on Improved Convolutional Neural Network for Face Recognition Security

—Face liveness detection is an important biometric authentication method for face recognition securitythat is used to determine a fake face from an authentic one. In this paper, a liveness detection method based on optimized LeNet5 is proposed. The LeNet-5 is optimized by increasing the convolution kerneland byintroducing a global average pooling. The simulation results show that the proposed model obtained the highest recognition rate of 99.95% as against the 96.67% and 98.23% accuracy from the Support Vector Machine (SVM) and LeNet-5 models, respectively.The results denote that the proposed model has a high recognition rate in face liveness detection

Face Anti-Spoofing Using Transformers With Relation-Aware Mechanism

Face anti-spoofing (FAS) is important to secure face recognition systems. Deep learning has obtained great success in this area, however, most existing approaches fail to consider comprehensive relation-aware local representations of live and spoof faces. To address this issue, we propose a Transformer-based Face Anti-Spoofing (TransFAS) model to explore comprehensive facial parts for FAS. Besides the multi-head self-attention which explores relations among local patches in the same layer, we propose cross-layer relation-aware attentions (CRA) to adaptively integrate local patches from different layers. Furthermore, to effectively fuse hierarchical features, we explore the best hierarchical feature fusion (HFF) structure, which can capture the complementary information between low-level artifacts and high-level semantic features for the spoofing patterns. With these novel modules, TransFAS not only improves the generalization capability of the classical vision transformer, but also achieves SOTA performance on multiple benchmarks, demonstrating the superiority of the transformer-based model for FAS.

Basic of Secrete Communication System for Military by Using ML and Steganography

Abstract: Information security is an important factor during transmitting secret information between two objects. Generally, we use cryptography for information hiding and sending secret messages in the form of text. Nowadays, there are several techniques used for hiding information in any medium. One such technique is steganography. In this technique, digital images are used for hiding information and the information is in the form of text, digital image, video or audio file may be used as a secret message. Using LSB Steganography Technique we can implement a high level of information security without any damage to the cover image. In this system we are using the hybrid approach i.e. cryptography andsteganography. So, our system has a higher security level than existing systems. With the development of machine learning, face recognition technology based on CNN (Convolutional Neural Network) has become the main method adopted in the field of face recognition security systems to securely access the confidential system. Keywords: Image data hiding, LSB Steganography, AES cryptography, Face Recognition, Machine Learning CNN technique

Design of Face Recognition Security System on Public Spaces

&lt;p class="Abstract"&gt;Implementation of physical security such as recruitment of security officers, installation of CCTV, and restrictions on public access have become commonplace nowadays. Computer systems equipped with archival storage media must be properly maintained, including computer systems containing sensitive information that must be stored in a locked and secure place. This study applies a security system using facial recognition to determine who is authorized over the data in the computer system—using Haar Cascade as a face detector and LBPH as a match between faces that can access and those that are not on the list. On the other hand, if the person is unidentified, preventive measures will be performed. Based on the result, the proposed system using Raspberry Pi 4 is able to identify a face using Haar Cascade algorithm with an accuracy of 68% and average duration process of 0.392s, and able to recognize face using LBPH algorithm with accuracy between 50.74% to 100% and average duration process of 0.548s.&lt;em&gt;&lt;/em&gt;&lt;/p&gt;

Hybrid HOG-SVM encrypted face detection and recognition model

Security plays a major role in an individual’s life to win this world with highly secure and authentic lifestyle with the digital equipment’s. The paper proposed an encryption based secure face detection and recognition model which can be implemented in daily life to generate a more robust and efficient security bubble around the world. The most crucial problem encountered during face recognition is due to the variation in face direction of an individual, the model solves the mentioned pose variation problem. The proposed model takes the help of face recognition library to recognize the face and use HOG (Histogram of Oriented Gradients) & SVM for checking the face authentication by performing an image match, the model also applies the concept of HOG to generate the encoded features from the image. The system is divided into two modules first is to detect a face and then match the detected face from the authentic persons dataset available. The system uses the concept of OpenCV library for giving a support system for the real time image. For data encryption, proposed model used the concept of DES3 and RSA algorithm. The proposed model gets 83.33% accuracy while tested for three different image types and states that the RSA algorithm performs encryption in less computational time.