Enhance Face Recognition Research Articles

Enhancing face recognition performance: a comprehensive evaluation of deep learning models and a novel ensemble approach with hyperparameter tuning

Enhancing face recognition performance: a comprehensive evaluation of deep learning models and a novel ensemble approach with hyperparameter tuning

FRCSyn-onGoing: Benchmarking and comprehensive evaluation of real and synthetic data to improve face recognition systems

This article presents FRCSyn-onGoing, an ongoing challenge for face recognition where researchers can easily benchmark their systems against the state of the art in an open common platform using large-scale public databases and standard experimental protocols. FRCSyn-onGoing is based on the Face Recognition Challenge in the Era of Synthetic Data (FRCSyn) organized at WACV 2024. This is the first face recognition international challenge aiming to explore the use of real and synthetic data independently, and also their fusion, in order to address existing limitations in the technology. Specifically, FRCSyn-onGoing targets concerns related to data privacy issues, demographic biases, generalization to unseen scenarios, and performance limitations in challenging scenarios, including significant age disparities between enrollment and testing, pose variations, and occlusions. To enhance face recognition performance, FRCSyn-onGoing strongly advocates for information fusion at various levels, starting from the input data, where a mix of real and synthetic domains is proposed for specific tasks of the challenge. Additionally, participating teams are allowed to fuse diverse networks within their proposed systems to improve the performance. In this article, we provide a comprehensive evaluation of the face recognition systems and results achieved so far in FRCSyn-onGoing. The results obtained in FRCSyn-onGoing, together with the proposed public ongoing benchmark, contribute significantly to the application of synthetic data to improve face recognition technology.

Enhancing Face Recognition with Detachable Self-Supervised Bypass Networks.

Attributed to the development of deep networks and abundant data, automatic face recognition (FR) has quickly reached human-level capacity in the past few years. However, the FR problem is not perfectly solved in case of large poses and uncontrolled occlusions. In this paper, we propose a novel bypass enhanced representation learning (BERL) method to improve face recognition under unconstrained scenarios. The proposed method integrates self-supervised learning and supervised learning together by attaching two auxiliary bypasses, a 3D reconstruction bypass and a blind inpainting bypass, to assist robust feature learning for face recognition. Among them, the 3D reconstruction bypass enforces the face recognition network to encode pose independent 3D facial information, which enhances the robustness to various poses. The blind inpainting bypass enforces the face recognition network to capture more facial context information for face inpainting, which enhances the robustness to occlusions. The whole framework is trained in end-to-end manner with two self-supervised tasks above and the classic supervised face identification task. During inference, the two auxiliary bypasses can be detached from the face recognition network, avoiding any additional computational overhead. Extensive experimental results on various face recognition benchmarks show that, without any cost of extra annotations and computations, our method outperforms state-of-the-art methods. Moreover, the learnt representations can also well generalize to other face-related downstream tasks such as the facial attribute recognition with limited labeled data.

A study of key issues in parallel algorithms for face recognition based on genetic neural networks

Abstract This study examines the effectiveness of Genetic Neural Networks (GNN) in face recognition, particularly in optimizing parallel algorithms to overcome the challenges posed by complex data. We have significantly improved recognition accuracy and computational efficiency by employing an adaptive genetic algorithm that fine-tunes neural network weights through Selection, crossover, and mutation. Our approach was tested across diverse datasets, covering variations in posture, age, ethnicity, and lighting conditions. The results demonstrate outstanding recognition rates: 99.82% on LFW, 97.94% on AgeDB-30, 95.11% on CFP-FP, 95.87% on CALFW, and 89.44% on CPLFW, showcasing exceptional robustness against complex lighting and occlusions. Additionally, our algorithm maintains balanced accuracy across different ethnicities with an overall recognition rate of 96.77% and boasts a substantial reduction in processing time to an average of 4.15 seconds. These advancements underscore the potential and practicality of our method in enhancing face recognition technology.

Enhancing face recognition with latent space data augmentation and facial posture reconstruction

The small amount of training data for many state-of-the-art deep learning-based Face Recognition (FR) systems causes a marked deterioration in their performance. Although a considerable amount of research has addressed this issue by inventing new data augmentation techniques, using either input space transformations or Generative Adversarial Networks (GAN) for feature space augmentations, these techniques have yet to satisfy expectations. In this paper, we propose an approach named Face Representation Augmentation (FRA) for augmenting face datasets. To the best of our knowledge, FRA is the first method that shifts its focus towards manipulating the face embeddings generated by any face representation learning algorithm to create new embeddings representing the same identity and facial emotion but with an altered posture. Extensive experiments conducted in this study convince us of the efficacy of the methodology and its power to provide noiseless, completely new facial representations to improve the training procedure of any FR algorithm. Therefore, FRA can help the recent state-of-the-art FR methods by providing more data for training FR systems. The proposed method, using experiments conducted on the Karolinska Directed Emotional Faces (KDEF) dataset, improves the identity classification accuracies by 9.52 %, 10.04 %, and 16.60 %, in comparison with the base models of MagFace, ArcFace, and CosFace, respectively.

Enhancing Face Recognition with Deep Learning Architectures: A Comprehensive Review

The progression of information discernment via facial identification and the emergence of innovative frameworks has exhibited remarkable strides in recent years. This phenomenon has been particularly pronounced within the realm of verifying individual credentials, a practice prominently harnessed by law enforcement agencies to advance the field of forensic science. A multitude of scholarly endeavors have been dedicated to the application of deep learning techniques within machine learning models. These endeavors aim to facilitate the extraction of distinctive features and subsequent classification, thereby elevating the precision of unique individual recognition. In the context of this scholarly inquiry, the focal point resides in the exploration of deep learning methodologies tailored for the realm of facial recognition and its subsequent matching processes. This exploration centers on the augmentation of accuracy through the meticulous process of training models with expansive datasets. Within the confines of this research paper, a comprehensive survey is conducted, encompassing an array of diverse strategies utilized in facial recognition. This survey, in turn, delves into the intricacies and challenges that underlie the intricate field of facial recognition within imagery analysis.

Enhancing Face Recognition Performance with Multispectral Imaging and Machine Learning: Comparison from Sift and Sift-Freak Feature Extraction

The development of modern societies faces many security and identification challenges. To meet this expectation, computer vision offers biometric solutions. Much research in recent years has focused on face recognition. Traditional facial recognition that uses color images has had many shortcomings, such as variation in illumination, smoke, rain, disguise, face concealment, makeup, etc. Light-insensitive infrared (IR) imaging is presented as an alternative to facial recognition in the visible to overcome the shortcomings of uncontrolled environments. However, IR also has weaknesses, such as facial occlusion by glasses, variation in body temperature, perfusion, etc. This paper proposes a new facial recognition architecture that uses several classification algorithms, detectors, and feature descriptors in multispectral imaging. A combination of SIFT and FREAK, feature extraction tools, was associated with classification algorithms such as SVM, logistic regression, and Random forest to conduct this study. Several experiments were made to evaluate the performance of the proposed recognition system. The validation process of the proposed multispectral face recognition method involved several important steps. First, experiments were carried out on visible and infrared spectrum images to measure the recognition system's performance. These experiments made it possible to compare the recognition performances between these two types of images. Then, fuse visible and infrared images were used to assess multispectral facial recognition. The goal was to maximize each spectrum's advantages while minimizing their disadvantages. Metrics were evaluated to measure the accuracy of the multispectral face recognition method. The performances were compared with classical facial recognition methods, such as facial recognition based on the visible spectrum or infrared imagery alone. The results showed that the proposed multispectral facial recognition method performed better than traditional methods, reaching a facial recognition score ranging from 76% to 95% in the IRIS database.

Enhancing Face Recognition Accuracy Using Data Pre-processing Method and YOLO

The recognition of objects is an essential aspect of visual perception and finds extensive usage in diverse fields such as self-driving vehicles, security, robotics, and image retrieval. In this study, we investigate the performance of the YOLOv5 (You Only Look Once) algorithm for object detection on the VOC2007 dataset. The YOLOv5 model achieved a moderate overall accuracy and precision, demonstrating its potential for object detection tasks. However, the performance varied across different categories, with lower accuracy observed for less frequent categories and difficulties in distinguishing between closely related categories. We identify potential improvements to the YOLOv5 model's performance, including class balancing using weighted sampling and data augmentation, which may help the model to better learn to detect objects from under-represented categories and improve its ability to distinguish between similar objects. The results of our study imply that the YOLO algorithm has potential for object detection and classification projects in computer vision, however further study and refinement are necessary to broaden its efficacy across a greater variety of object classes and real-world scenarios.

Exploring intervention strategies to enhance face recognition abilities in adolescents with autism spectrum disorder: assessing the impact of verbalization and Navon tasks

PurposeThe purpose of this study is to examine the effectiveness of two intervention strategies – identifying the global letter of Navon letters and verbally describing a face – in improving face recognition abilities of adolescents with autism spectrum disorder (ASD).Design/methodology/approachA total of 15 adolescents with ASD and 15 typically developing (TD) adolescents participated in this study, which consisted of three experimental conditions: verbalization, Navon and control. Participants were presented with a target face for a short period of time, then completed a task specific to the condition in a counterbalanced order, followed by a face recognition task.FindingsThe results of this study showed no significant interaction effects between groups and conditions, either in accuracy or reaction time. However, TD adolescents performed better than ASD adolescents in terms of accuracy. In addition, the verbalization condition resulted in better performance with slower response times compared to the control condition.Originality/valueThese findings suggest that verbalization may be a useful strategy to enhance face recognition abilities in ASD and TD individuals.

Optimized Collaborative Brain-Computer Interfaces for Enhancing Face Recognition.

The aim of this study is to maximize group decision performance by optimally adapting EEG confidence decoders to the group composition. We train linear support vector machines to estimate the decision confidence of human participants from their EEG activity. We then simulate groups of different size and membership by combining individual decisions using a weighted majority rule. The weights assigned to each participant in the group are chosen solving a small-dimension, mixed, integer linear programming problem, where we maximize the group performance on the training set. We therefore introduce optimized collaborative brain-computer interfaces (BCIs), where the decisions of each team member are weighted according to both the individual neural activity and the group composition. We validate this approach on a face recognition task undertaken by 10 human participants. The results show that optimal collaborative BCIs significantly enhance team performance over other BCIs, while improving fairness within the group. This research paves the way for practical applications of collaborative BCIs to realistic scenarios characterized by stable teams, where optimizing the decision policy of a single group may lead to significant long-term benefits of team dynamics.

FACE RECOGNITION WITH HYBRID TECHNIQUES

Face recognition framework is still in test by numerous applications particularly in close perception and in security frameworks. Generally all utilizations of face recognition utilize enormous information sets, making challenges in present time preparing and effectiveness. This paper contains a structure to enhance face recognition framework which have a few phases. For good result in face recognition framework a few upgrades are critical at each stage. A novel plan is displayed in this paper which gives the better execution for face recognition framework. This plan incorporates expanding in datasets, particularly huge datasets which are required for profound learning. Changing the picture differentiate proportion and pivoting the picture at a few edges which can enhance the recognition precision. At that point, trimming the proper territory of face for highlight extraction and getting the best element vector for face recognition finally. The last after effect of this plan will demonstrate that the given structure is able for distinguishing and perceiving faces with various postures, foundations, and appearance in genuine or present time.

Performance Analysis of Face Detection system using HOG and QualHOG Features

Inspired by the expansion of minimal effort advanced cameras in cell phones being conveyed in computerized systems, we think about the connection between perceptual picture quality and an exemplary PC vision errand of face recognition. We measure the corruption in execution of a well known and compelling face detector when human-saw image quality is corrupted by twists usually happening in catch, stockpiling, and transmission of facial pictures, including clamor, obscure, and pressure. It is observed that, inside a certain scope of picture quality, an unobtrusive increment in picture quality can radically enhance face recognition execution. These outcomes can be utilized to guide asset or transfer speed distribution in securing or correspondence/conveyance frameworks that are connected with face location undertakings. In this work a perceptual quality QualHOG feature is used. Face locators prepared on these new components give measurably huge change in resilience to picture bends over a solid gauge. Distortion dependent which is more distorted uninformed variations of the face indicators are proposed and assessed on a huge database of face pictures speaking to an extensive variety of mutilations. A one-sided variation of the preparing calculation is additionally recommended that further improves the power of these face locators. To encourage this exploration, we have developed another dataset in our lab for further study.

Cyborg groups enhance face recognition in crowded environments.

Recognizing a person in a crowded environment is a challenging, yet critical, visual-search task for both humans and machine-vision algorithms. This paper explores the possibility of combining a residual neural network (ResNet), brain-computer interfaces (BCIs) and human participants to create “cyborgs” that improve decision making. Human participants and a ResNet undertook the same face-recognition experiment. BCIs were used to decode the decision confidence of humans from their EEG signals. Different types of cyborg groups were created, including either only humans (with or without the BCI) or groups of humans and the ResNet. Cyborg groups decisions were obtained weighing individual decisions by confidence estimates. Results show that groups of cyborgs are significantly more accurate (up to 35%) than the ResNet, the average participant, and equally-sized groups of humans not assisted by technology. These results suggest that melding humans, BCI, and machine-vision technology could significantly improve decision-making in realistic scenarios.

Joint Face Alignment and 3D Face Reconstruction with Application to Face Recognition.

Face alignment and 3D face reconstruction are traditionally accomplished as separated tasks. By exploring the strong correlation between 2D landmarks and 3D shapes, in contrast, we propose a joint face alignment and 3D face reconstruction method to simultaneously solve these two problems for 2D face images of arbitrary poses and expressions. This method, based on a summation model of 3D faces and cascaded regression in 2D and 3D shape spaces, iteratively and alternately applies two cascaded regressors, one for updating 2D landmarks and the other for 3D shape. The 3D shape and the landmarks are correlated via a 3D-to-2D mapping matrix, which is updated in each iteration to refine the location and visibility of 2D landmarks. Unlike existing methods, the proposed method can fully automatically generate both pose-and-expression-normalized (PEN) and expressive 3D faces and localize both visible and invisible 2D landmarks. Based on the PEN 3D faces, we devise a method to enhance face recognition accuracy across poses and expressions. Both linear and nonlinear implementations of the proposed method are presented and evaluated in this paper. Extensive experiments show that the proposed method can achieve the state-of-the-art accuracy in both face alignment and 3D face reconstruction, and benefit face recognition owing to its reconstructed PEN 3D face.

A Novel Feature Extraction Approach Using Principal Component Analysis and Quantum behaved Particle Swarm Optimization-Support Vector Networks for Enhancing Face Recognition

A Novel Feature Extraction Approach Using Principal Component Analysis and Quantum behaved Particle Swarm Optimization-Support Vector Networks for Enhancing Face Recognition

Enhancing Face Recognition from Massive Weakly Labeled Data of New Domains

Training data are critical in face recognition systems. Labeling a large scale dataset for a particular domain needs lots of manpower. Without dataset related to current face recognition domain, we can’t get a strong face recognition model with existing public datasets. In this paper, we propose a semi-supervised method to automatically construct strong dataset which can be trained to achieve better performance on the target domain from massive weakly labeled data. In the case of Asian face recognition, a well trained VRCN model by CASIA, which achieves 98.63% on LFW and 91.76% on YTF, only achieves 88.53% recognition rate on our test dataset of Asian faces. We collect 530,560 weakly labeled Asian face images of 7962 identities, and get a cleaned dataset of size 285,933. Model trained by the cleaned dataset with VRCN network and same strategy achieves 95.33% recognition rate on the Asian face test dataset (6.8% improved).

Multiplex image representation for enhanced recognition

A multiscale approach to exploiting existing image descriptors (LBP and HOG) is proposed recently in order to enhance face recognition performance (Ubiquitous computing and ambient intelligence. Personalisation and user adapted services. Springer, 532–539, 2014) and (A multiscale method for HOG-based face and palmprint recognition. Technical report, Ulster University, 2015), where multiple single-sourced, spatially-varied feature vectors at different scales are calculated from images and then fused through an image distance function. This multiscale approach has led to significant improvements in face recognition over the single scale approach. In this paper we present an analysis of this multiscale approach from feature engineering perspective and evaluation result for the image distance function on palmprint recognition, thus providing an insight into and also extending the applicability of this approach. We also present a new method of utilising these spatially-varied feature vectors from an image—joining these feature vectors head to tail to form a larger feature vector which is used as a multiplex representation of the image. Such an image representation can then be used by any vector-based feature extraction and classification algorithms. This representation scheme is evaluated experimentally in face recognition, and the results show this scheme is competitive to the distance-based method having the additional advantage of being usable in a wider range of machine learning algorithms. The main contributions of this paper are (1) an insight into this multiscale approach to utilising existing image descriptors such as LBP and HOG; (2) a new method of using these multiple feature vectors; and (3) extension of the multiscale approach to palmprint recognition.

Does attention to low spatial frequencies enhance face recognition? An individual differences approach

Does attention to low spatial frequencies enhance face recognition? An individual differences approach

Holistic face training enhances face processing in developmental prosopagnosia

Prosopagnosia has largely been regarded as an untreatable disorder. However, recent case studies using cognitive training have shown that it is possible to enhance face recognition abilities in individuals with developmental prosopagnosia. Our goal was to determine if this approach could be effective in a larger population of developmental prosopagnosics. We trained 24 developmental prosopagnosics using a 3-week online face-training program targeting holistic face processing. Twelve subjects with developmental prosopagnosia were assessed before and after training, and the other 12 were assessed before and after a waiting period, they then performed the training, and were then assessed again. The assessments included measures of front-view face discrimination, face discrimination with view-point changes, measures of holistic face processing, and a 5-day diary to quantify potential real-world improvements. Compared with the waiting period, developmental prosopagnosics showed moderate but significant overall training-related improvements on measures of front-view face discrimination. Those who reached the more difficult levels of training ('better' trainees) showed the strongest improvements in front-view face discrimination and showed significantly increased holistic face processing to the point of being similar to that of unimpaired control subjects. Despite challenges in characterizing developmental prosopagnosics' everyday face recognition and potential biases in self-report, results also showed modest but consistent self-reported diary improvements. In summary, we demonstrate that by using cognitive training that targets holistic processing, it is possible to enhance face perception across a group of developmental prosopagnosics and further suggest that those who improved the most on the training task received the greatest benefits.

The effect of gaze direction on three-dimensional face recognition in infant brain activity

In three-dimensional face recognition studies, it is well known that viewing rotating faces enhance face recognition. For infants, our previous study indicated that 8-month-old infants showed recognition of three-dimensional rotating faces with a direct gaze, and they did not learn with an averted gaze. This suggests that gaze direction may affect three-dimensional face recognition in infants. In this experiment, we used near-infrared spectroscopy to measure infants' hemodynamic responses to averted gaze and direct gaze. We hypothesized that infants would show different neural activity for averted and direct gazes. The responses were compared with the baseline activation during the presentation of non-face objects. We found that the concentration of oxyhemoglobin increased in the temporal cortex on both sides only during the presentation of averted gaze compared with that of the baseline period. This is the first study to show that infants' brain activity in three-dimensional face processing is different between averted gaze and direct gaze.