Face Identification Research Articles

Working memory capacity modulates Serial dependence in facial Identity: Evidence from behavioral and EEG data.

Serial dependence (SD) is said to occur when the judgment of a current stimulus is drawn toward a no longer relevant stimulus from the recent past. Working memory (WM) contributes to the ability to discriminate between irrelevant and relevant sensory impressions. How WM contributes to SD in facial identity remains to be fully understood. In the present study, we used a WM task with varying load interleaved with a facial matching task to investigate the relations between SD in facial identity and WM. EEG was recorded to capture perceptual face processing and WM maintenance. Overall, statistically significant SD effects were obtained only at high WM load. In addition, individual differences in WM capacity predicted SD. Participants with low WM capacity showed significant SD effects regardless of changes in WM load, while participants with high WM capacity showed significant SD only at high WM load. EEG results showed significant differences in perceptual and WM processing for participants with high WM capacity, while no significant differences were obtained for participants with low WM capacity. Neural correlates of WM maintenance, and perceptual processing of a previous face were directly related to SD. The results suggest that WM capacity contributes to SD in facial identity, in terms of selective suppression of task-irrelevant and selective maintenance of task-relevant information.

Human-like face pareidolia emerges in deep neural networks optimized for face and object recognition.

The human visual system possesses a remarkable ability to detect and process faces across diverse contexts, including the phenomenon of face pareidolia--seeing faces in inanimate objects. Despite extensive research, it remains unclear why the visual system employs such broadly tuned face detection capabilities. We hypothesized that face pareidolia results from the visual system's optimization for recognizing both faces and objects. To test this hypothesis, we used task-optimized deep convolutional neural networks (CNNs) and evaluated their alignment with human behavioral signatures and neural responses, measured via magnetoencephalography (MEG), related to pareidolia processing. Specifically, we trained CNNs on tasks involving combinations of face identification, face detection, object categorization, and object detection. Using representational similarity analysis, we found that CNNs that included object categorization in their training tasks represented pareidolia faces, real faces, and matched objects more similarly to neural responses than those that did not. Although these CNNs showed similar overall alignment with neural data, a closer examination of their internal representations revealed that specific training tasks had distinct effects on how pareidolia faces were represented across layers. Finally, interpretability methods revealed that only a CNN trained for both face identification and object categorization relied on face-like features-such as 'eyes'-to classify pareidolia stimuli as faces, mirroring findings in human perception. Our results suggest that human-like face pareidolia may emerge from the visual system's optimization for face identification within the context of generalized object categorization.

Learnable Anchor Embedding for Asymmetric Face Recognition

Face verification and identification traditionally follow a symmetric matching approach, where the same model (e.g., ResNet-50 vs. ResNet-50) generates embeddings for both gallery and query images, ensuring compatibility. However, real-world scenarios often demand asymmetric matching, especially when query devices have limited computational resources or employ heterogeneous models (e.g., ResNet-50 vs. SwinTransformer). This asymmetry can degrade face recognition performance due to incompatibility between embeddings from different models. To tackle this asymmetric face recognition problem, we introduce the Learnable Anchor Embedding (LAE) model, which features two key innovations: the Shared Learnable Anchor and a Light Cross-Attention Mechanism. The Shared Learnable Anchor is a dynamic attractor, aligning heterogeneous gallery and query embeddings within a unified embedding space. The Light Cross-Attention Mechanism complements this alignment process by reweighting embeddings relative to the anchor, efficiently refining their alignment within the unified space. Extensive evaluations of several facial benchmark datasets demonstrate LAE’s superior performance, particularly in asymmetric settings. Its robustness and scalability make it an effective solution for real-world applications such as edge-device authentication, cross-platform verification, and environments with resource constraints.

Exploring the Challenges of Aadhaar based Face Recognition in Unrestricted Environments

- Technology improvements have resulted in a high criminal rate, which has raised serious concerns about security-related issues. Face recognition can be used to identify people because every person tends to possess a particular attribute. One of the main applications of facial recognition is in the field of video surveillance. To reduce the increasing criminal rates, this technology is responsible for extracting features from the human face and further identifying them. The CCTV footage might also be used to identify suspects at a crime scene. However, the criminals would be found by recognizing and comparing the suspect's face to the database repository that would include data on people based on their Aadhaar card. Aadhaar photos can be used to identify and recognize faces in a facial recognition system since Aadhaar is a unique identification system in India and mandatory for every Indian resident. However, there are many challenges faced by Aadhaar based face recognition. Face recognition in uncontrolled environments like public surveillance faces significant challenges such as low-resolution images, pose variations, and masked faces. This paper highlights the challenges faced by Aadhaar enabled face recognition in surveillance scenes. This article aims to provide a brief survey of the Aadhaar-enabled face recognition system for criminal identification and recognizing individuals. There is limited work with the Aadhaar dataset due to privacy and security reasons. This article also emphasizes the difficulties that will be encountered in implementing facial recognition as well as strategies for enhancing it while taking various tradeoffs into account. It also recommends opportunities for further study in the application of facial identification in other fields. Key Words: Aadhaar, Surveillance, Face recognition, Criminals

Assessing bias and computational efficiency in vision transformers using early exits

Face recognition with deep learning is generally approached as a problem of capacity. The field has seen progressively deeper, more complex models or larger, more highly variant data sets. The data sets can be problematic, as they are often scraped indiscriminately from the internet. This results in an uncertain, and often heavily unbalanced distribution of race, gender, age and other aspects of the subjects, which is then manifested in the decisions of the models trained on them. The carbon footprint of machine learning is a concern. A real push is developing to reduce the energy consumption of machine learning as we strive for a more eco-friendly society. In addition, due to many instances of misuse by law enforcement and other agencies, unbiased models for face recognition are now fundamental to the practical application of the field. We present an approach using the state of the art Vision Transformer and Early Exits for reducing compute budget without significantly affecting performance. We develop a system for face recognition and identification with a closed-set gallery and show that with a small reduction in performance, a reasonable reduction in compute cost can be obtained using our method. Second, we investigate how these early Exits interact with the bias model through a robust evaluation of matching scores on a racially balanced data set. We show that matching scores vary heavily between cohorts, and these variations are magnified at the earlier exits.

Brain connectivity for constructing new face representations in typical adults versus a prosopagnosic patient.

Will our brains get to know a new face better if we look at its external features first? Here we offer neurophysiological evidence of the relevance of external versus internal facial features for constructing new face representations, by contrasting successful face processing with a prototypical case of face agnosia. A woman with acquired prosopagnosia (E.C.) and 14 age-matched typical participants (7 women) were exposed to a face-feature matching task. External (E), internal (I) features, and whole target faces of unknown individuals (from an IdentiKit gallery) were displayed according to two different sequences: E →I→whole faces, or I→E→whole faces. Then, we studied the induced EEG activity using 'isolated effective coherence' to analyse the intracortical causal information flow among face-sensitive nodes. Initial presentation of external features (E before I), when compared to internal ones, triggered connections encompassing extensively the right-hemisphere face processing pathway [from posterior visual cortices for initial structural analysis, towards both intermediate (occipitotemporal) and high-level (prefrontal) relay stations], in which face-identity is thought to emerge progressively. Also, whereas exposure to internal features as second stimulus seemed to demand some sort of basic visual processing, external features triggered again more widespread and integrative connections. Connections for whole faces closing the E-I sequence resembled those for external features initiating the same sequence. Meanwhile, the predominant connections for whole faces completing the I-E sequence were more restricted to specific brain areas, with relevant prefrontal activity and a few connected nodes in right posterior regions, suggesting high attentional load plus initial and intermediate processing of face identity. Interestingly, the pattern of connections that distinguished typical participants from E.C. in the I-E sequence was the recruitment of left posterior visual regions, presumably underlying analytical subroutines for structural encoding of facial stimuli. These findings support that initial exposure to external features, followed by internal ones, provides the best visual cue to acquire new face configurations. Nevertheless, in case of face agnosia after right posterior damage, relying preferentially on internal features and left hemisphere specialized subroutines might be an alternative for cognitive training.

Rapport Management Framework in Classroom Interaction: A Socio-pragmatic Analysis of Teachers' Utterances

This qualitative descriptive study employing socio-pragmatic analysis aimed to analyze teachers' utterances using Spencer-Oatey's Rapport Management Framework, essential for improving teacher-student relationships and enhancing classroom dynamics. A 6-hour classroom observation and interview with 10 teachers were conducted to collect the data. The framework includes face and social rights. In terms of face, it involved the quality face, which comprised well-being concern, dialogue encouragement, openness to diverse perspectives, respect for diverse opinions, feedback and encouragement, fairness and humility; social identity face, which comprised of behavioral guidance, authority establishment, group engagement, effort encouragement, participation seeking, attention direction, guidance of understanding, correctness assurance, and support reinforcement; and relational face, which comprised of managing classroom dynamics, gentle correction, commitment to personalized support, promotion of mutual support, validation and clarification strategy. On the other hand, sociality rights involved association and equity rights: the latter consisted of the principles of involvement, empathy, and respect, while the former consisted of cost-benefit consideration, fairness and reciprocity, and autonomy-control. The results imply that both face and sociality rights can be sources of offense. Face is important for maintaining students' personal/relational/social value, while sociality rights help ensure fairness, consideration and behavioral appropriateness. These elements are important for teachers when interacting with students because they promote respectful communication and strengthen teacher-student relationships. The inclusion of these concepts in the curriculum helps create a positive and inclusive classroom environment.

Robust Face Detection and Identification under Occlusion using MTCNN and RESNET50

In today's rapidly evolving world, where technology is progressing swiftly, there is an increasing demand for facial recognition systems. Technologies are similar to digital forensics in that they can recognize people by scanning faces. However, one key problem they confront is dealing with covered or occluded faces, which might restrict recognition of faces in real-world situations. To overcome this issue, we created a system that is capable of identifying individuals even when their faces are veiled. We used the face detector algorithm called Multi-Task Cascaded Convolutional Neural Network (MTCNN) for face detection with 99.8% accuracy. Further we have conducted feature extraction and pre-processing on our self-created dataset. Our project utilizes the power of deep learning model: Residual Network (ResNet50), the form of deep neural network architectures well-suited for the job of features extraction. These features are matched by using Cosine similarity with accuracy of 92%. By leveraging the capabilities in the deep learning algorithms, this project provides a robust solution for automating the recognition of partially occluded faces.

Designing Face Detection Systems with Gray Wolf Optimization

The main objective of this paper project was to create a state-of-the-art face identification technique that can handle the various difficulties caused by changes in illumination, occlusions, and facial emotions. Face detection is a cornerstone of computer vision, facilitating diverse applications ranging from surveillance systems to human-computer interaction. Throughout this paper, the comprehensive exploration of advancing face detection methodologies has been undertaken, culminating in developing and evaluating a novel approach. The challenges posed by variations in facial expressions, lighting conditions, and occlusions necessitated a multifaceted solution. Our proposed method, which consists of interconnected steps, works quite well to overcome these challenges. Using deep learning architectures to increase feature extraction and discrimination was beneficial in the initial stage of fine-tuning Residual Networks (ResNet-50) to serve as the Region-based Convolutional Neural Network (Faster R-CNN) framework classifier. The process of gradually optimizing thresholds, such as batch size, learning rate, and detection threshold, involved using the Gray Wolf optimization technique (GWO). The conversion process was accelerated and improved overall detection process efficiency and accuracy using a clever fusion of machine learning and metaheuristic optimization techniques. A key component of our methodology is the careful data processing, which was necessary to ensure. The suggested method was carefully examined on a particular dataset, and the 94% training accuracy that was attained together with an identical test dataset accuracy highlights the method's resilience. These findings support the effectiveness of our approach in reducing false positives and negatives, resulting in unmatched recall and precision in the detection system. The discovery has significant significance as it can potentially improve face detection systems' performance and reliability in various real-world applications, such as human-computer interaction and surveillance. Convolutional neural networks, deep learning architectures, and metaheuristic optimization approaches were synergized to produce a new and reliable solution.

Study on the Impact of LDA Preprocessing on Pig Face Identification with SVM.

In this study, the implementation of traditional machine learning models in the intelligent management of swine is explored, focusing on the impact of LDA preprocessing on pig facial recognition using an SVM. Through experimental analysis, the kernel functions for two testing protocols, one utilizing an SVM exclusively and the other employing a combination of LDA and an SVM, were identified as polynomial and RBF, both with coefficients of 0.03. Individual identification tests conducted on 10 pigs demonstrated that the enhanced protocol improved identification accuracy from 83.66% to 86.30%. Additionally, the training and testing durations were reduced to 0.7% and 0.3% of the original times, respectively. These findings suggest that LDA preprocessing significantly enhances the efficiency of individual pig identification using an SVM, providing empirical evidence for the deployment of SVM classifiers in mobile and embedded systems.

Strengthening Security in Clouds Through Cloud Computing Authentication Using Facial Image Forensics

Numerous cyber threats may succeed in cloud platforms owing to unproductive authentication methods. Multiple credential authentication is a vital precautionary measure that helps reinforce cloud security for warding off imminent data breaches and illegal access. This kind of authentication approach strongly guarantees that trustworthy clients are only ratified to get cloud services, making it less tiresome for clients and more secure for organizations. This study proposes an Image Forensics-based Dual Credential Authentication (IF-DCA) based on username (i.e., preferably email ID to avoid replica) and client photographs as password. Along with a username, a rapid and acquainted user action of capturing a photograph using a web camera is adequate for this method; hence it does not need any expensive, special hardware devices. During the registration phase, along with the username, the user's face image is stored on the cloud server as a password. When registered clients need to access cloud resources or data, they should log in on the server with their username and their photograph. To improve the authentication approach, this work proposes an Attribute-controlled Conditional Generative Adversarial Network (ACC-GAN) to generate face images of the same user at various age groups. ACC-GAN includes an additional attribute control unit and an age prediction unit to synthesize photo-realistic face images with aging effects. After matching the username and face image with the registered username and ACC-GAN-generated images, the face identification module provides access to the authorized cloud user. The performance of IF-DCA is assessed through the Cross-Age Celebrity Dataset using the MATLAB R2018b/deep learning toolbox. The empirical analysis reveals that the ACC-GAN achieves better performance measures such as 98.40% accuracy, 99.3% sensitivity, 90.0% specificity, and 95.7% precision. The performance of IF-DCA is analyzed using the verification time for username, user image, and complete authentication against the number of users. Received: 4 September 2024 | Revised: 14 October 2024 | Accepted: 7 November 2024 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The Cross-Age Celebrity Datasets that support the findings of this study are openly available at https://paperswithcode.com/dataset/cacd. Author Contribution Statement Pranali Dahiwal: Conceptualization, Software, Formal analysis, Resources, Writing – original draft, Visualization, Project administration. Vijay Khare: Methodology, Validation, Investigation, Data curation, Writing – review & editing, Supervision.

Neural correlates of facial recognition deficits in autism spectrum disorder: a comprehensive review.

Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by significant impairments in social interaction, often manifested in facial recognition deficits. These deficits hinder individuals with ASD from recognizing facial identities and interpreting emotions, further complicating social communication. This review explores the neural mechanisms underlying these deficits, focusing on both functional anomalies and anatomical differences in key brain regions such as the fusiform gyrus (FG), amygdala, superior temporal sulcus (STS), and prefrontal cortex (PFC). It has been found that the reduced activation in the FG and atypical activation of the amygdala and STS contribute to difficulties in processing facial cues, while increased reliance on the PFC for facial recognition tasks imposes a cognitive load. Additionally, disrupted functional and structural connectivity between these regions further exacerbates facial recognition challenges. Future research should emphasize longitudinal, multimodal neuroimaging approaches to better understand developmental trajectories and design personalized interventions, leveraging AI and machine learning to optimize therapeutic outcomes for individuals with ASD.

Research trends in livestock facial identification – A review

Research trends in livestock facial identification – A review

The role of expansion and adaptability of face-space for individual differences in face identity processing.

Individuals can strongly vary in their ability to process face identity. Understanding the mechanisms driving these differences is important for theoretical development, and in clinical and applied contexts. Here we investigate the role of face-space properties in relation to individual face identity processing skills. We consider two fundamental properties of face-space: expansion (how distant from each other similar faces are located in such space) and adaptability (the degree to which these distances change over time). Fifty-two participants performed a face detection task, with faces systematically varying in their location in face-space, and a comprehensive face identity processing test battery. We replicate previous results indicating a detection advantage for typical, as compared with distinctive faces. Critically, we find that neither our measure of face-space expansion nor that of face-space adaptability are related to individual face processing abilities. While future studies might benefit from the use of more sensitive measures of face-space properties, these results suggest that the two examined here do not contribute to individual differences in face processing abilities as previous studies suggest.

IOT BASED DOOR ACCESS CONTROL SYSTEM USING ESP32CAM

The home security system has become very important for each house. Previously, most doors are open by victimization ancient ways, resembling keys, security cards, Arcanum or pattern. However, incidents such as a key loss has crystal rectifier to a lot of worrying cases such as theft and identity fraud. This has become a major issue. during this paper we've got given a wise Wi-Fi Door Lock mistreatment the ESP32 CAM associate degreed wire app. during this easy operating model, once someone is detected by a PIR motion sensing element an ESP32CAM captures the face then telegram sends the notifications on phone, so the owner offers unlock/lock commands within the telegram. It mechanically responds to it command. Mainly, this project is for security purpose reception and industries. Anyone these days cares about security, whether or not it's information security or the protection of their own home. Digital door locks have mature quite prevailing in recent years as technology has advanced and therefore the use of IOT has increased. A digital lock doesn't need a physical key to operate, instead counting on Radio-Frequency Identification (RFID), fingerprint, Face ID, pins, passwords, and alternative strategies to try to do so.

View-symmetric representations of faces in human and artificial neural networks.

View symmetry has been suggested to be an important intermediate representation between view-specific and view-invariant representations of faces in the human brain. Here, we compared view-symmetry in humans and a deep convolutional neural network (DCNN) trained to recognise faces. First, we compared the output of the DCNN to head rotations in yaw (left-right), pitch (up-down) and roll (in-plane rotation). For yaw, an initial view-specific representation was evident in the convolutional layers, but a view-symmetric representation emerged in the fully-connected layers. Consistent with a role in the recognition of faces, we found that view-symmetric responses to yaw were greater for same identity compared to different identity faces. In contrast, we did not find a similar transition from view-specific to view-symmetric representations in the DCNN for either pitch or roll. These findings suggest that view-symmetry emerges when opposite rotations of the head lead to mirror images. Next, we compared the view-symmetric patterns of response to yaw in the DCNN with corresponding behavioural and neural responses in humans. We found that responses in the fully-connected layers of the DCNN correlated with judgements of perceptual similarity and with the responses of higher visual regions. These findings suggest that view-symmetric representations may be computationally efficient way to represent faces in humans and artificial neural networks for the recognition of identity.

Exposing the Forgery Clues of DeepFakes via Exploring the Inconsistent Expression Cues

The pervasive prevalence of DeepFakes poses a profound threat to individual privacy and the stability of society. Believing the synthetic videos of a celebrity and trumping up impersonated forgery videos as authentic are just a few consequences generated by DeepFakes. We investigate current detectors that blindly deploy deep learning techniques that are not effective in capturing subtle clues of forgery when generative models produce remarkably realistic faces. Inspired by the fact that synthetic operations inevitably modify the regions of eyes and mouth to match the target face with the identity or expression of the source face, we conjecture that the continuity of facial movement patterns representing expressions that existed in the veritable faces will be disrupted or completely broken in synthetic faces, making it a potentially formidable indicator for DeepFake detection. To prove this conjecture, we utilize a dual‐branch network to capture the inconsistent patterns of facial movements within eyes and mouth regions separately. Extensive experiments on popular FaceForensics++, Celeb‐DF‐v1, Celeb‐DF‐v2, and DFDC‐Preview datasets have demonstrated not only effectiveness but also the robust capability of our method to outperform the state‐of‐the‐art baselines. Moreover, this work represents greater robustness against adversarial attacks, achieving ASR of 54.8% in the I‐FGSM attack and 43.1% in the PGD attack on the DeepFakes dataset of FaceForensics++, respectively.

Face Identification based on different Convolutional Neural Network Models for Security System: A Review

Face Identification based on different Convolutional Neural Network Models for Security System: A Review

Audio-Visual Contrastive Pre-train for Face Forgery Detection

The highly realistic avatar in the metaverse may lead to deepfakes of facial identity. Malicious users can more easily obtain the three-dimensional structure of faces, thus using deepfake technology to create counterfeit videos with higher realism. To automatically discern facial videos forged with the advancing generation techniques, deepfake detectors need to achieve stronger generalization abilities. Inspired by transfer learning, neural networks pre-trained on other large-scale face-related tasks would provide fundamental features for deepfake detection. We propose a video-level deepfake detection method based on a temporal transformer with a self-supervised audio–visual contrastive learning approach for pre-training the deepfake detector. The proposed method learns motion representations in the mouth region by encouraging the paired video and audio representations to be close while unpaired ones to be diverse. The deepfake detector adopts the pre-trained weights and partially fine-tunes on deepfake datasets. Extensive experiments show that our self-supervised pre-training method can effectively improve the accuracy and robustness of our deepfake detection model without extra human efforts. Compared with existing deepfake detection methods, our proposed method achieves better generalization ability in cross-dataset evaluations.

Evaluating and Enhancing Face Anti-Spoofing Algorithms for Light Makeup: A General Detection Approach.

Makeup modifies facial textures and colors, impacting the precision of face anti-spoofing systems. Many individuals opt for light makeup in their daily lives, which generally does not hinder face identity recognition. However, current research in face anti-spoofing often neglects the influence of light makeup on facial feature recognition, notably the absence of publicly accessible datasets featuring light makeup faces. If these instances are incorrectly flagged as fraudulent by face anti-spoofing systems, it could lead to user inconvenience. In response, we develop a face anti-spoofing database that includes light makeup faces and establishes a criterion for determining light makeup to select appropriate data. Building on this foundation, we assess multiple established face anti-spoofing algorithms using the newly created database. Our findings reveal that the majority of these algorithms experience a decrease in performance when faced with light makeup faces. Consequently, this paper introduces a general face anti-spoofing algorithm specifically designed for light makeup faces, which includes a makeup augmentation module, a batch channel normalization module, a backbone network updated via the Exponential Moving Average (EMA) method, an asymmetric virtual triplet loss module, and a nearest neighbor supervised contrastive module. The experimental outcomes confirm that the proposed algorithm exhibits superior detection capabilities when handling light makeup faces.