Search

Face recognition is used in many applications such as access control, automobile security, criminal identification, immigration, healthcare, cyber security, and so on. Each person has his/her own unique face, so the face can help distinguish people from each other. Feature extraction process plays a fundamental role in accuracy of face recognition, and many algorithms have been presented to extract more informative features from the face image. In this paper, an efficient texture descriptor is proposed based on local information of the face image. In the proposed method, at first, face image is split into several sub-images in such a way that each sub-image includes one of the facial parts such as eyes, nose, and lips. Second, texture features are extracted from each sub-image using a new local pattern descriptor, and then features of sub-images are concatenated to construct feature vector. Finally, the face image is compared to images in a dataset based on a similarity measure. In addition, particle swarm optimization algorithm is used to assign weight to the features of different parts of the face image. To evaluate the proposed algorithm, four face datasets, Yale, ORL, GT and KDEF, are used. Implementation results show that the proposed method outperforms recent methods in terms of accuracy, receiver operating characteristic (ROC) curve, and area under ROC curve. [ABSTRACT FROM AUTHOR]

This study evaluates the performance of the Viola Jones and YOLOv3 algorithms for facial recognition under different conditions and highlights their strengths and weaknesses. Analysis focusses on facial emotions, angle recognition, lighting, and the effects of hidden facial features. YOLOv3 outperformed the Viola-Jones algorithm in angle-based recognition with more robustness. Both algorithms performed exceptionally well in different lighting conditions, with 100% recognition rates in artificial, natural, high-contrast, and dark surroundings. This shows that they are highly adaptive to changing lighting conditions. When individual facial characteristics, such as the forehead or eyes, were concealed, the Viola-Jones algorithm showed excellent reliability. When the nose and eyes were concealed, however, its performance dropped to 77%. YOLOv3, on the other hand, consistently achieved a 100% recognition rate, indicating that it handled inadequate facial data better, even in scenarios where multiple significant attributes were concealed. Both algorithms proven their resistance to dynamic face changes by achieving 100% recognition rates over a wide range of expressions and proving that facial expressions had no effect on their recognition accuracy. These algorithms should be improved in the future for extreme angles and partial occlusions, and their integration with other recognition methods should be investigated.

An efficient face recognition system is essential for security and authentication-based applications. However, real-time face recognition systems have a few significant concerns, including face pose orientations. In the last decade, numerous solutions have been introduced to estimate distinct face pose orientations. Nevertheless, these solutions must be adequately addressed for the three main face pose orientations: Yaw, Pitch, and Roll. This paper proposed a novel deep transfer learning-based multitasking approach for solving three integrated tasks, i.e., face detection, landmarks detection, and face pose estimation. The face pose variation vulnerability has been intensely investigated here underlying three modules: image preprocessing, feature extraction module through deep transfer learning, and regression module for estimating the face poses. The experiments are performed on the well-known benchmark dataset Annotated Faces in the Wild (AFW). We evaluate the outcomes of the experiments to reveal that our proposed approach is superior to other recently available solutions.

Abstract The human face is commonly used for identity verification. While this task was once exclusively performed by humans, technological advancements have seen automated facial recognition systems (AFRS) integrated into many identification scenarios. Although many state-of-the-art AFRS are exceptionally accurate, they often require human oversight or involvement, such that a human operator actions the final decision. Previously, we have shown that on average, humans assisted by a simulated AFRS (sAFRS) failed to reach the level of accuracy achieved by the same sAFRS alone, due to overturning the system’s correct decisions and/or failing to correct sAFRS errors. The aim of the current study was to investigate whether participants’ trust in automation was related to their performance on a one-to-one face matching task when assisted by a sAFRS. Participants (n = 160) completed a standard face matching task in two phases: an unassisted baseline phase, and an assisted phase where they were shown the identification decision (95% accurate) made by a sAFRS prior to submitting their own decision. While most participants improved with sAFRS assistance, those with greater relative trust in automation achieved larger gains in performance. However, the average aided performance of participants still failed to reach that of the sAFRS alone, regardless of trust status. Nonetheless, further analysis revealed a small sample of participants who achieved 100% accuracy when aided by the sAFRS. Our results speak to the importance of considering individual differences when selecting employees for roles requiring human–algorithm interaction, including identity verification tasks that incorporate facial recognition technologies.

In recent years, the sample mining strategy has been integrated into the loss function of face recognition, significantly improving the performance of face recognition. But most of the work focuses on how to mine difficult samples during the training phase, without considering the potential unrecognized sample images in the difficult samples, resulting in poor recognition performance of the model for low-quality facial images. To solve this problem, this paper proposes a hybrid adaptive loss function MixFace that combines sample difficulty adaptation and image quality adaptation. The loss function combines the CurricularFace based on curriculum learning with the image adaptive loss function AdaFace. The feature norm is incorporated into the loss function as an image quality indicator. On the premise of focusing on image quality, this paper focuses on simple samples in the early training stage and difficult samples in the later training stage, reducing the network model’s attention to some low-quality unrecognized samples in difficult samples. Trained on CASIA-WebFace and MS1MV2 datasets, MixFace shows varying degrees of performance improvement compared with CurricularFace and AdaFace on high-quality test sets LFW, CFP_FP, AgeDB, CALFW, and CPLFW. At the same time, MixFace shows better recognition performance than CurricularFace and AdaFace on medium quality test sets IJB-B, IJB-C and low-quality test set TinyFace. Experimental results show that MixFace can effectively reduce the interference of unrecognized images, thereby improving the performance of low-quality face recognition. At the same time, benefiting from the curriculum learning method in MixFace, it can still maintain good performance for high-quality face recognition.

Face recognition uses biometric technologies to identify humans based on their facial characteristics. This method is commonly used to restrict information access control. The benefits of face recognition systems encompass their ease of use and security. The human face recognition process consists of face detection, face tracking, and face recognition. The process uses some algorithms: the Viola-Jones for face detection, the Kanade-Lucas-Tomasi (KLT) for face tracking, and the principal component analysis (PCA) for face recognition. Furthermore, this research proposed face recognition with an encryption process to protect the data stored in a database. The encryption process consists of two main processes: confusion and diffusion. The confusion process is randomizing the position of the original image pixels. This research utilized the Arnold’s cat map (ACM) for the confusion process, and the diffusion process was performed using the XOR operation with the key generated from the 1D chaos system. Three different 1D chaos systems, namely logistic map, Bernoulli map, and tent map, were compared to see which chaos system had the best encryption results. Tests were conducted by comparing various parameters on the three proposed 1D chaos systems, including correlation coefficient, histogram analysis, entropy value, number of pixel rate changes (NPCR), and unified average change intensity (UACI). Based on testing the image encryption results, the diffusion process utilizing the tent map produced the best image encryption compared to other chaotic systems.

This article describes and analyzes the development of a system for identifying and recognizing the cognitive behavior of students to determine interest in facial expressions. The purpose of the study is to find suitable technologies for the implementation of this system. The definition of emotions will allow organizing control over the quality of the educational process, conducting statistics on the cognitive behavior of students during classes, and showing the level of interest of students in the material presented. The identification system will automatically determine and register the time of arrival and departure of students in real time. Based on the joint application of the Viola – Jones method and the nearest neighbors method using histograms of centrally symmetric local binary images, a system for face recognition in a real-time video sequence has been developed. The structure of the project is described and the software is developed in the Python programming language using the Keras open-source library. The developed system consists of two subsystems: an identification system and a cognitive behavior recognition system. The scientific novelty lies in an integrated approach to the development and research of algorithms for real-time face recognition and identification for solving applied problems.

The human face is commonly used for identity verification. While this task was once exclusively performed by humans, technological advancements have seen automated facial recognition systems (AFRS) integrated into many identification scenarios. Although many state-of-the-art AFRS are exceptionally accurate, they often require human oversight or involvement, such that a human operator actions the final decision. Previously, we have shown that on average, humans assisted by a simulated AFRS (sAFRS) failed to reach the level of accuracy achieved by the same sAFRS alone, due to overturning the system's correct decisions and/or failing to correct sAFRS errors. The aim of the current study was to investigate whether participants' trust in automation was related to their performance on a one-to-one face matching task when assisted by a sAFRS. Participants (n = 160) completed a standard face matching task in two phases: an unassisted baseline phase, and an assisted phase where they were shown the identification decision (95% accurate) made by a sAFRS prior to submitting their own decision. While most participants improved with sAFRS assistance, those with greater relative trust in automation achieved larger gains in performance. However, the average aided performance of participants still failed to reach that of the sAFRS alone, regardless of trust status. Nonetheless, further analysis revealed a small sample of participants who achieved 100% accuracy when aided by the sAFRS. Our results speak to the importance of considering individual differences when selecting employees for roles requiring human–algorithm interaction, including identity verification tasks that incorporate facial recognition technologies. [ABSTRACT FROM AUTHOR]

Face recognition is one of the most important research topics in computer vision. Indeed, the face is an important means of communication with humans and it is highly needed for daily contact. Face recognition technology is applied in many biometric applications such as security, video surveillance, access control systems, and forensics. In this technology, hashing has recently made encouraging progress due to its fast retrieval speed and low storage cost. In this work, we propose an effective face recognition framework based on hashing functions. It attempts to leverage a cascaded architecture with two stages of analyzing different visual information based on image hashing. Specifically, we first introduce a filter to overlook a large number of dissimilar identities in terms of local visual information. Similar identities are found quickly through random independent hash functions inspired by Locality Sensitive Hashing (LSH). Next, we further refine candidates and recognize the most similar identities according to global visual information. The global feature is obtained by hashing each face into a high-quality binary feature space using Discrete Cosine Transform (DCT) coefficients. The proposed method is evaluated on three well-known and one combined face dataset. The obtained results, and the provided face recognition application program, demonstrate that the proposed framework improves the recognition rate and significantly reduces recognition time. [ABSTRACT FROM AUTHOR]

It has been established that convolutional neural networks are susceptible to elaborate tiny universal adversarial perturbations (UAPs) in natural image classification tasks. However, UAP attacks against face recognition systems have not been fully explored. This paper proposes a spatial perturbation method that generates UAPs with local stealthiness by learning variable flow field to fine-tune facial key regions (KRT-FUAP). We ensure that the generated adversarial perturbations are positioned within reasonable regions of the face by designing a mask specifically tailored to facial key regions. In addition, we pay special attention to improving the effectiveness of the attack while maintaining the stealthiness of the perturbation and achieve the dual optimization of aggressiveness and stealthiness by accurately controlling the balance between adversarial loss and stealthiness loss. Experiments conducted on the frameworks of IResNet50 and MobileFaceNet demonstrate that our proposed method achieves an attack performance comparable to existing natural image universal attack methods, but with significantly improved stealthiness. [ABSTRACT FROM AUTHOR]

With the increasing use of smart homes and IoT devices, security has become a significant concern. This paper presents a method to enhance smart home security using Deep Convolutional Neural Networks (DCNN) and multiple cameras. In this approach, three cameras are used to capture images from different angles, and these images are analysed using DCNNs including VGG16, VGG19 and DenseNet to detect potential intruders. Known for their excellent performance in image classification, these DCNN models aim to improve the accuracy and efficiency of the security system, thereby reducing false alarms and missed detections. Additionally, the system allows authorized individuals to remotely disable the security system, increasing convenience and usability. The proposed method has shown significant improvement in human presence detection and facial recognition, achieving 99.79% accuracy in classifying home occupants and intruders. This performance is superior to alternative models such as SVM, KNN, and complex decision trees. This paper introduces a new method that integrates multiple cameras with DCNN to boost the performance of security systems. [ABSTRACT FROM AUTHOR]

Recently, face recognition based on homomorphic encryption for privacy preservation has garnered significant attention. However, there are two major challenges with homomorphic encryption methods: the security and efficiency of face recognition systems. We present a more efficient and secure PUM (Privacy preserving security Using Multi‐key homomorphic encryption) mechanism for facial recognition. By integrating feature grouping with parallel computing, we enhance the efficiency of homomorphic operations. The use of multi‐key encryption ensures the security of the facial recognition system. This approach improves the security and speed of facial recognition systems in cloud computing scenarios, increasing the original 128‐bit security to a maximum of 1664‐bit security. In terms of efficiency, comparing encrypted images takes only 0.302 s, with an accuracy rate of 99.425%. When applied to a campus scenario, the average search time for a facial template library containing 700 encrypted features is approximately 1.5 s. Consequently, our solution not only ensures user privacy but also demonstrates superior operational efficiency and practical value. In comparison to recently emerged ciphertext facial recognition systems, our solution has demonstrated notable enhancements in both security and time efficiency. [ABSTRACT FROM AUTHOR]

Identity-independent factors, such as variations of pose, expression, illumination, etc., are the key challenges in face recognition. To avoid the effects of these factors, existing face recognition methods usually adopt two approaches: pose-invariant face feature extracting and face normalization before feature extraction. Contrary to these, we propose a single deep model jointly performing face normalization and representation learning tasks for face recognition, named normalization and reconstruction general adversarial network (NRGAN). First, the unified NRGAN model can boost the performance of the two tasks for each other. Second, NRGAN can synthesize normalized face images without the requirement of paired data, which makes our method have better generalization ability to the uncontrolled environment. Third, a factor-invariant identity disentanglement training strategy is introduced to decouple the identity feature representation from other factors without using any of these factors’ labels. Extensive experiment results on four currently popular face datasets demonstrate the effectiveness of NRGAN on both normalized face synthesis and face recognition tasks. [ABSTRACT FROM AUTHOR]

With the introduction of deep learning, a significant amount of research has been conducted in the field of computer vision in the past decade. In particular, research on object detection (OD) continues to progress rapidly. However, despite these advances, some limitations need to be overcome to enable real-world applications of deep learning-based OD models. One such limitation is inaccurate OD when image quality is poor or a target object is small. The performance degradation phenomenon for small objects is similar to the fundamental limitations of an OD model, such as the constraint of the receptive field, which is a difficult problem to solve using only an OD model. Therefore, OD performance can be hindered by low image quality or small target objects. To address this issue, this study investigates the compatibility of super-resolution (SR) and OD techniques to improve detection, particularly for small objects. We analyze the combination of SR and OD models, classifying them based on architectural characteristics. The experimental results show a substantial improvement when integrating OD detectors with SR models. Overall, it was demonstrated that, when the evaluation metrics (PSNR, SSIM) of the SR models are high, the performance in OD is correspondingly high as well. Especially, evaluations on the MS COCO dataset reveal that the enhancement rate for small objects is 9.4% higher compared to all objects. This work provides an analysis of SR and OD model compatibility, demonstrating the potential benefits of their synergistic combination. The experimental code can be found on our GitHub repository. [ABSTRACT FROM AUTHOR]

Face recognition is important for both visual and social cognition. While prosopagnosia or face blindness has been known for seven decades and face-specific neurons for half a century, the molecular genetic mechanism is not clear. Here we report results after 17 years of research with classic genetics and modern genomics. From a large family with 18 congenital prosopagnosia (CP) members with obvious difficulties in face recognition in daily life, we uncovered a fully cosegregating private mutation in the MCTP2 gene which encodes a calcium binding transmembrane protein expressed in the brain. After screening through cohorts of 6589, we found more CPs and their families, allowing detection of more CP associated mutations in MCTP2. Face recognition differences were detected between 14 carriers with the frameshift mutation S80fs in MCTP2 and 19 noncarrying volunteers. Six families including one with 10 members showed the S80fs-CP correlation. Functional magnetic resonance imaging found association of impaired recognition of individual faces by MCTP2 mutant CPs with reduced repetition suppression to repeated facial identities in the right fusiform face area. Our results have revealed genetic predisposition of MCTP2 mutations in CP, 76 years after the initial report of prosopagnosia and 47 years after the report of the first CP. This is the first time a gene required for a higher form of visual social cognition was found in humans. [ABSTRACT FROM AUTHOR]

This work presents a novel method for facial recognition that aggregates weak facial parts to overcome the problem of recognizing faces with incomplete information. By applying a landmark detection method that approximates 70 facial landmarks, we divide the face into 12 weak information parts. These parts are further refined using superpixel segmentation, which increases the precision of face feature analysis. A weak classifier is training a weak part in order to recognize the face. We then aggregate these weak classifiers with a late-fusion voting technique to build a robust classification system. Our technique has been examined on subsets of the FaceScrub and VGG datasets, with a particular focus on frontal face pictures. Experiments show the effectiveness of our method, with an accuracy range of 54.5593% to 91.6346% on the tiny FaceScrub dataset and 45.8654% to 76.4231% on the tiny VGG dataset, depending on how many weak information parts are considered. This result demonstrates the potential for combining available weak information parts to improve the accuracy of face recognition systems, especially in situations where conventional approaches may be hampered by missing or occluded facial data. [ABSTRACT FROM AUTHOR]

Emotional stimuli and events are better and more easily remembered than neutral ones. However, this advantage appears to come at a cost, namely a decreased accuracy for peripheral, emotion-irrelevant details. There is some evidence, particularly in the visual modality, that this trade-off also applies to emotional expressions, leading to a difficulty in identifying an unfamiliar individual's identity when presented with an expression different from the one encountered at encoding. On the other hand, past research also suggests that identity recognition memory benefits from exposure to different encoding exemplars, although whether this is also the case for emotional expressions, particularly voices, remains unknown. Here, we directly addressed these questions by conducting a series of voice and face identity memory online studies, using a within-subject old/new recognition test in separate unimodal modules. In the Main Study, half of the identities were encoded with four presentations of one single expression (angry, fearful, happy, or sad; Uni condition) and the other half with one presentation of each emotion (Multi condition); all identities, intermixed with an equal number of new ones, were presented with a neutral expression in a subsequent recognition test. Participants (N = 547, 481 female) were randomly assigned to one of four groups in which a different Uni single emotion was used. Results, using linear mixed models on response choice and drift-diffusion-model parameters, revealed that high-arousal expressions interfered with emotion-independent identity recognition accuracy, but that such deficit could be compensated by presenting the same individual with various expressions (i.e., high exemplar variability). These findings were confirmed by a significant correlation between memory performance and stimulus arousal, across modalities and emotions, and by two follow-up studies (Study 1: N = 172, 150 female; Study 2: N = 174, 154 female), which extended the original observations and ruled out some potential confounding effects. Taken together, the findings reported here expand and refine our current knowledge of the influence of emotion on memory, and highlight the importance of, and interaction between, exemplar variability and emotional arousal in identity recognition memory. [ABSTRACT FROM AUTHOR]

While subjective methods like the Yanagihara system and the House-Brackmann system are standard in evaluating facial paralysis, they are limited by intra- and inter-observer variability. Meanwhile, quantitative objective methods such as electroneurography and electromyography are time-consuming. Our aim was to introduce a swift, objective, and quantitative method for evaluating facial movements. We developed an application software (app) that utilizes the facial recognition functionality of the iPhone (Apple Inc., Cupertino, USA) for facial movement evaluation. This app leverages the phone's front camera, infrared radiation, and infrared camera to provide detailed three-dimensional facial topology. It quantitatively compares left and right facial movements by region and displays the movement ratio of the affected side to the opposite side. Evaluations using the app were conducted on both normal and facial palsy subjects and were compared with conventional methods. Our app provided an intuitive user experience, completing evaluations in under a minute, and thus proving practical for regular use. Its evaluation scores correlated highly with the Yanagihara system, the House-Brackmann system, and electromyography. Furthermore, the app outperformed conventional methods in assessing detailed facial movements. Our novel iPhone app offers a valuable tool for the comprehensive and efficient evaluation of facial palsy. [ABSTRACT FROM AUTHOR]

We address the need for a large-scale database of children's faces by using generative adversarial networks (GANs) and face-age progression (FAP) models to synthesize a realistic dataset referred to as "HDA-SynChildFaces". Hence, we proposed a processing pipeline that initially utilizes StyleGAN3 to sample adult subjects, which is subsequently progressed to children of varying ages using InterFaceGAN. Intra-subject variations, such as facial expression and pose, are created by further manipulating the subjects in their latent space. Additionally, this pipeline allows the even distribution of the races of subjects, allowing the generation of a balanced and fair dataset with respect to race distribution. The resulting HDA-SynChildFaces consists of 1,652 subjects and 188,328 images, each subject being present at various ages and with many different intra-subject variations. We then evaluated the performance of various facial recognition systems on the generated database and compared the results of adults and children at different ages. The study reveals that children consistently performworse than adults on all tested systems and that the degradation in performance is proportional to age. Additionally, our study uncovers some biases in the recognition systems, with Asian and black subjects and females performing worse than white and Latino-Hispanic subjects and males. [ABSTRACT FROM AUTHOR]

Copyright of Przegląd Elektrotechniczny is the property of Przeglad Elektrotechniczny and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Nowadays, biometric systems have emerged as a powerful tool for personal identification. Advanced research with significant results has been provided. Despite the important progress, there is a big need for improvement in the performance of security applications. More recently, an Autonomous Learning Multi-Model Classifier of 0- Order (ALMMo-0) has been proposed as a universal efficient tool of classification, autonomous, non-iterative, and fully explainable solving the problem of supervised pattern recognition. Thereby, our aim with this paper is to propose a new efficient methodology based on the ALMMo-0 classifier for authentication systems exploiting face modality. The proposed methodology is entirely data-driven, non-iterative, and feedforward. The proposed approach extracts the most relevant features from the face image by the Gabor filter bank descriptor, which is then fed into the ALMMo-0 classifier that extracts automatically the data clouds and builds its multimodal structure, forms its AnYa Fuzzy rule base (FRB) sub-classifiers for each class, generating objectively the score of confidence based on the mutual distribution then classify the new data using "winner takes all" strategy, as a result, the system decides whether the person is genuine or an imposter. Strong evidence of ALMMo-0 was found when experiments were conducted on nine face databases. Results were presented in the form of rank-1, equal error rate (EER), cumulative match curve (CMC), and receiver operating characteristic (ROC) curves. Furthermore, to provide more valuable information about our proposed system, results were also presented in the form of True Positive Rate (TPR) or the Genuine Acceptance Rate (GAR). The results demonstrated high performance of the proposed approach with not just a low error rate (EER) of 0.0% and a high accuracy (rank-1) of 100% but also high explainability and low computational complexity. [ABSTRACT FROM AUTHOR]

This paper mainly introduces related research of the full screen technology for face recognition under OLED screen. By analyzing and studying the influence of infrared light irradiation on OLED screen, especially the influence of infrared light irradiation on TFT, the relevant mechanism analysis is completed, and the corresponding optimization solution is proposed, so as to realize face recognition under the screen, achieve the perfect technical combination of full screen and infrared face recognition, and realize true full screen display. [ABSTRACT FROM AUTHOR]

Face is one of the most important biometric traits utilized by humans for recognition. Face recognition is the prominent biometric method for human authentication, and it is used in several domains due to its unique features, non-intrusive, and convenience compared to other biometric systems like fingerprint or palmprint scans. Although the field of face recognition has advanced significantly, there are still problems that prevent accuracy from surpassing that of humans. This study proposes a novel and effective framework, named Face Identification utilizing Spider Hierarchy with a Classic Classifier (FISH-CC), aimed at recognizing a person's face, gender, and age. This framework incorporates a novel face boundary localization scheme based on cooperative game theory (CGT), enhancing facial detection performance by accurately detecting facial contour. Features are extracted from the detected faces using a modified local binary pattern (mLBP). To optimize feature selection, a CGT-based algorithm, known as the extended contribution selection algorithm (ECSA) with forward feature selection (FFS), is implemented. Finally, Spider Hierarchy (SH) integrated with a Classic Classifier (CC) is used for face identification. To assess the effectiveness of the proposed method, a number of tests are carried out, and the labeled faces in the wild (LFW) database are utilized to validate the performance. The outcomes of this study demonstrated that the proposed FISH-CC achieves a superior accuracy rate of 99.60% when compared to the existing approaches. [ABSTRACT FROM AUTHOR]