Face Data Research Articles

Kapsül Ağları Kullanılarak Gerçek ve Sahte Yüz Verilerinin Sınıflandırılması

Recently, with the advancement of technology, artificial intelligence has begun to be used in many areas. It is used in many fields such as artificial intelligence, image processing, natural language processing, suggestion systems. The increase in the use of artificial intelligence has revealed the need for data. This situation has led to the production of new data from existing data. Today, generative adversarial networks (GAN) synthesize a wide variety of data inspired by existing data. These fake data produced from real image data can sometimes cause undesirable situations. It is an important issue to know whether the images, which are especially important for security, are fake or not. In this study, the classification of real human face data and fake face data generate from these data has been made. Fake face data were generated with the help of StyleGAN2-ADA from a small-sized dataset created by collecting facial data of a famous person. It is aimed to classify the generated fake face data and real face data with the capsule network model.

Face Duplication Identifier Using Artificial Nerves

The facial recognition system develops a basic identity verification system based on the natural features of human faces. The study included duplicate passport identification, which checks each person's facial accuracy through a sample of facial data. The data used in this study were 180 face samples at the training stage and 30 face samples at the testing stage. The face sample taken is a forward-facing face that is not obstructed by an object. Face image recognition in this study combines GLCM method, color moment, shape extraction and backpropagation algorithm. The test process recognition rate is 78.83%.

PhenoScore quantifies phenotypic variation for rare genetic diseases by combining facial analysis with other clinical features using a machine-learning framework.

Several molecular and phenotypic algorithms exist that establish genotype-phenotype correlations, including facial recognition tools. However, no unified framework that investigates both facial data and other phenotypic data directly from individuals exists. We developed PhenoScore: an open-source, artificial intelligence-based phenomics framework, combining facial recognition technology with Human Phenotype Ontology data analysis to quantify phenotypic similarity. Here we show PhenoScore's ability to recognize distinct phenotypic entities by establishing recognizable phenotypes for 37 of 40 investigated syndromes against clinical features observed in individuals with other neurodevelopmental disorders and show it is an improvement on existing approaches. PhenoScore provides predictions for individuals with variants of unknown significance and enables sophisticated genotype-phenotype studies by testing hypotheses on possible phenotypic (sub)groups. PhenoScore confirmed previously known phenotypic subgroups caused by variants in the same gene for SATB1, SETBP1 and DEAF1 and provides objective clinical evidence for two distinct ADNP-related phenotypes, already established functionally.

Detecting at-risk mental states for psychosis (ARMS) using machine learning ensembles and facial features.

Our study aimed to develop a machine learning ensemble to distinguish "at-risk mental states for psychosis" (ARMS) subjects from control individuals from the general population based on facial data extracted from video-recordings. 58 non-help-seeking medication-naïve ARMS and 70 healthy subjects were screened from a general population sample. At-risk status was assessed with the Structured Interview for Prodromal Syndromes (SIPS), and "Subject's Overview" section was filmed (5-10min). Several features were extracted, e.g., eye and mouth aspect ratio, Euler angles, coordinates from 51 facial landmarks. This elicited 649 facial features, which were further selected using Gradient Boosting Machines (AdaBoost combined with Random Forests). Data was split in 70/30 for training, and Monte Carlo cross validation was used. Final model reached 83% of mean F1-score, and balanced accuracy of 85%. Mean area under the curve for the receiver operator curve classifier was 93%. Convergent validity testing showed that two features included in the model were significantly correlated with Avolition (SIPS N2 item) and expression of emotion (SIPS N3 item). Our model capitalized on short video-recordings from individuals recruited from the general population, effectively distinguishing between ARMS and controls. Results are encouraging for large-screening purposes in low-resource settings.

Distilling knowledge for occlusion robust monocular 3D face reconstruction

Recently, there have been significant advancements in the 3D face reconstruction field, largely driven by monocular image-based deep learning methods. However, these methods still face challenges in reliable deployments due to their sensitivity to facial occlusions and inability to maintain identity consistency across different occlusions within the same facial image. To address these issues, we propose two frameworks: DistillationAssistedMonoImageOcclusionRobustification (DAMIOR) and DuplicateImagesAssistedMultiOcclusionsRobustification (DIAMOR). The DAMIOR framework leverages the knowledge from the OcclusionFrailTrainer (OFT) network to enhance robustness against facial occlusions. Our proposed method overcomes the sensitivity to occlusions and improves reconstruction accuracy. To tackle the issue of identity inconsistency, the DIAMOR framework utilizes the estimates from DAMIOR to mitigate inconsistencies in geometry and texture, collectively known as identity, of the reconstructed 3D faces. We evaluate the performance of DAMIOR on two variations of the CelebA test dataset: empirical occlusions and irrational occlusions. Furthermore, we analyze the performance of the proposed DIAMOR framework using the irrational occlusion-based variant of the CelebA test dataset. Our methods outperform state-of-the-art approaches by a significant margin. For example, DAMIOR reduces the 3D vertex-based shape error by 41.1% and the texture error by 21.8% for empirical occlusions. Besides, for facial data with irrational occlusions, DIAMOR achieves a substantial decrease in shape error by 42.5% and texture error by 30.5%. These results demonstrate the effectiveness of our proposed methods.

4D facial analysis: A survey of datasets, algorithms and applications

Facial information plays an important role in human communication, e.g., rich and nuanced facial expressions effectively convey emotions. Traditional data such as 2D facial images and videos are susceptible to perturbation from lighting and occlusion, while 3D static data such as mesh models lack the temporal information which is necessary to describe facial dynamics. Therefore, researches on 4D facial data (3D facial models together with time as the fourth dimension) have received considerable attention in recent years. 4D data can simultaneously reflect the complex facial temporal and space information. To fully explore these characteristics, we present a systematic overview of the 4D facial research in this paper. We give a review in terms of historical development of 4D facial datasets, acquisition process of these datasets, related algorithms and applications, and discussion of outstanding issues. We also analyze the 4D facial research works by summarizing and comparing them; in particular, we present the results on three kinds of tasks conducted on the 4D facial datasets, i.e., expression-related tasks (AU detection, expression recognition and retrieval), generation tasks (3D facial reconstruction, facial expression and facial animation generation), and other tasks (facial registration, facial recognition and facial disease diagnosis). Finally, we summarize the future research directions.

Willingness of sharing facial data for emotion recognition: a case study in the insurance market

Willingness of sharing facial data for emotion recognition: a case study in the insurance market

Preliminary study on three-dimensional morphological reconstruction method for external nose defect based on three-dimensional face template

Objective: To provide a new solution for the digital design of nasal prostheses, this study explores the three-dimensional (3D) facial morphology completion method for external nasal defects based on the non-rigid registration process of 3D face template. Methods: A total of 20 male patients with tooth defect and dentition defect who visited the Department of Prosthodontics, Peking University School and Hospital of Stomatology from June to December 2022 were selected, age 18-45 years old. The original 3D facial data of patients were collected, and the 3D facial data of the external nose defect was constructed in Geomagic Wrap 2021 software. Using the structured 3D face template data constructed in the previous research of the research group, the 3D face template was deformed and registered to the 3D facial data of external nose defect (based on the morphology of non-defective area) by non-rigid registration algorithm (MeshMonk program), and the personalized deformed data of the 3D face template was obtained, as the complemented facial 3D data. Based on the defect boundary of the 3D facial data of the external nose defect, the complemented external nose 3D data can be cut out from the complemented facial 3D data. Then the nasofacial angle and nasolabial angle of the complemented facial 3D data and the original 3D facial data was compared and analyzed, the ratio between the nose length and mid-face height, nose width and medial canthal distance of the complemented facial 3D data was measured, the edge fit between the edge curve of the complemented external nose 3D data and the defect edge curve of the 3D facial data of external nose defect was evaluated, and the morphological difference of the nose between the complemented external nose 3D data and the original 3D facial data was analyzed. Results: There was no significant statistically difference (t=-0.23, P=0.823; Z=-1.72, P=0.086) in the nasofacial angle (28.2°±2.9°, 28.4°±3.5° respectively) and nasolabial angle [95.4°(19.2°), 99.9°(9.5°) respectively] between the 20 original 3D facial data and the complemented facial 3D data. The value of the ratio of nose length to mid-face height in the complemented facial 3D data was 0.63±0.03, and the value of the ratio of nose width to medial canthal distance was 1.07±0.08. The curve deviation (root mean square value) between the edge curve of the complemented external nose 3D data and the defect edge curve of the 3D facial data of external nose defect was (0.37±0.09) mm, the maximum deviation was (1.14±0.32) mm, and the proportion of the curve deviation value within±1 mm was (97±3)%. The distance of corresponding nose landmarks between the complemented facial 3D data and the original 3D facial data were respectively, Nasion: [1.52(1.92)] mm; Pronasale: (3.27±1.21) mm; Subnasale: (1.99±1.09) mm; Right Alare: (2.64±1.34) mm; Left Alare: (2.42± 1.38) mm. Conclusions: The method of 3D facial morphology completion of external nose defect proposed in this study has good feasibility. The constructed complemented external nose 3D data has good facial coordination and edge fit, and the morphology is close to the nose morphology of the original 3D facial data.

Identifying personal physiological data risks to the Internet of Everything: the case of facial data breach risks

Personal physiological data is the digital representation of physical features that identify individuals in the Internet of Everything environment. Such data includes characteristics of uniqueness, identification, replicability, irreversibility of damage, and relevance of information, and this data can be collected, shared, and used in a wide range of applications. As facial recognition technology has become prevalent and smarter over time, facial data associated with critical personal information poses a potential security and privacy risk of being leaked in the Internet of Everything application platform. However, current research has not identified a systematic and effective method for identifying these risks. Thus, in this study, we adopted the fault tree analysis method to identify risks. Based on the risks identified, we then listed intermediate events and basic events according to the causal logic, and drew a complete fault tree diagram of facial data breaches. The study determined that personal factors, data management and supervision absence are the three intermediate events. Furthermore, the lack of laws and regulations and the immaturity of facial recognition technology are the two major basic events leading to facial data breaches. We anticipate that this study will explain the manageability and traceability of personal physiological data during its lifecycle. In addition, this study contributes to an understanding of what risks physiological data faces in order to inform individuals of how to manage their data carefully and to guide management parties on how to formulate robust policies and regulations that can ensure data security.

Analysis of candidate genes for cleft lip ± cleft palate using murine single-cell expression data.

Introduction: Cleft lip ± cleft palate (CL/P) is one of the most common birth defects. Although research has identified multiple genetic risk loci for different types of CL/P (i.e., syndromic or non-syndromic forms), determining the respective causal genes and understanding the relevant functional networks remain challenging. The recent introduction of single-cell RNA sequencing (scRNA-seq) has provided novel opportunities to study gene expression patterns at cellular resolution. The aims of our study were to: (i) aggregate available scRNA-seq data from embryonic mice and provide this as a resource for the craniofacial community; and (ii) demonstrate the value of these data in terms of the investigation of the gene expression patterns of CL/P candidate genes. Methods and Results: First, two published scRNA-seq data sets from embryonic mice were re-processed, i.e., data representing the murine time period of craniofacial development: (i) facial data from embryonic day (E) E11.5; and (ii) whole embryo data from E9.5-E13.5 from the Mouse Organogenesis Cell Atlas (MOCA). Marker gene expression analyses demonstrated that at E11.5, the facial data were a high-resolution representation of the MOCA data. Using CL/P candidate gene lists, distinct groups of genes with specific expression patterns were identified. Among others we identified that a co-expression network including Irf6, Grhl3 and Tfap2a in the periderm, while it was limited to Irf6 and Tfap2a in palatal epithelia, cells of the ectodermal surface, and basal cells at the fusion zone. The analyses also demonstrated that additional CL/P candidate genes (e.g., Tpm1, Arid3b, Ctnnd1, and Wnt3) were exclusively expressed in Irf6+ facial epithelial cells (i.e., as opposed to Irf6- epithelial cells). The MOCA data set was finally used to investigate differences in expression profiles for candidate genes underlying different types of CL/P. These analyses showed that syndromic CL/P genes (syCL/P) were expressed in significantly more cell types than non-syndromic CL/P candidate genes (nsCL/P). Discussion: The present study illustrates how scRNA-seq data can empower research on craniofacial development and disease.

A novel facial expression recognition algorithm using geometry β –skeleton in fusion based on deep CNN

Facial expression recognition (FER) methods based on single-source facial data often suffer from reduced accuracy or unpredictability due to facial occlusion or illumination changes. To address this, a new technique called Fusion-CNN is proposed. It improves accuracy by extracting hybrid features using a β-skeleton undirected graph and an ellipse with parameters trained using a 1D-CNN. In addition, a 2D-CNN is trained on the same image. The outputs from these two subnetworks are fused, and their features are concatenated to create a feature vector for classification in a deep neural network. The proposed method is evaluated on four public face datasets: the extended Cohn-Kanade (CK+) dataset, the Japanese Female Facial Expression (JAFFE) dataset, Karolinska Directed Emotional Faces (KDEF), and Oulu-CASIA. The experimental results show that Fusion-CNN outperforms other algorithms, achieving recognition accuracy of 98.22%, 93.07%, 90.30%, and 90.13% for the CK+, JAFFE, KDEF, and Oulu-CASIA datasets, respectively.

HUMAN GENDER DETECTION SYSTEM BASED ON FACIAL IMAGE USING CONVOLUTIONAL NEURAL NETWORK ALGORITHM

The demand for system automation has been continuously increasing with the current technological developments. One of these advancements is in the implementation of face recognition. Camera capabilities have evolved from merely capturing images or videos to being able to process the resulting images. Facial images contain a wealth of information, one of which is the gender information of the individuals. To obtain this information, facial image classification using deep learning is required. In this scientific paper, the author utilizes the Convolutional Neural Network algorithm implemented with the Python programming language and employs TensorFlow as its framework. The research aims to predict human gender based on facial images. The dataset used in this study is obtained from the kaggle.com dataset provider, consisting of 9,600 male facial data and 9,600 female facial data. The data is divided into a training and testing set, with an 80% ratio for training data and a 20% ratio for testing data from the total available data. The model training process is performed for 15 epochs with 768 steps in each epoch. The testing results show that the Convolutional Neural Network method achieves a validation accuracy of approximately 91%. The developed program runs well through a webcam.

Comparison Study of Extraction Accuracy of 3D Facial Anatomical Landmarks Based on Non-Rigid Registration of Face Template

(1) Background: Three-dimensional (3D) facial anatomical landmarks are the premise and foundation of facial morphology analysis. At present, there is no ideal automatic determination method for 3D facial anatomical landmarks. This research aims to realize the automatic determination of 3D facial anatomical landmarks based on the non-rigid registration algorithm developed by our research team and to evaluate its landmark localization accuracy. (2) Methods: A 3D facial scanner, Face Scan, was used to collect 3D facial data of 20 adult males without significant facial deformities. Using the radial basis function optimized non-rigid registration algorithm, TH-OCR, developed by our research team (experimental group: TH group) and the non-rigid registration algorithm, MeshMonk (control group: MM group), a 3D face template constructed in our previous research was deformed and registered to each participant's data. The automatic determination of 3D facial anatomical landmarks was realized according to the index of 32 facial anatomical landmarks determined on the 3D face template. Considering these 32 facial anatomical landmarks manually selected by experts on the 3D facial data as the gold standard, the distance between the automatically determined and the corresponding manually selected facial anatomical landmarks was calculated as the "landmark localization error" to evaluate the effect and feasibility of the automatic determination method (template method). (3) Results: The mean landmark localization error of all facial anatomical landmarks in the TH and MM groups was 2.34 ± 1.76 mm and 2.16 ± 1.97 mm, respectively. The automatic determination of the anatomical landmarks in the middle face was better than that in the upper and lower face in both groups. Further, the automatic determination of anatomical landmarks in the center of the face was better than in the marginal part. (4) Conclusions: In this study, the automatic determination of 3D facial anatomical landmarks was realized based on non-rigid registration algorithms. There is no significant difference in the automatic landmark localization accuracy between the TH-OCR algorithm and the MeshMonk algorithm, and both can meet the needs of oral clinical applications to a certain extent.

Designing a Facial Expression Detection System to Determine the Level of Customer Satisfaction with K-Nearest Neighbor Method

Facial expressions are one of the ways humans communicate to convey one's emotions to their communication partner nonverbally. Therefore, human facial expressions can be used for various purposes, one of which is knowing customer satisfaction. So far, customers of Bank Rakyat Indonesia (BRI) provide feedback on service quality using only a polling system, namely by filling out a criticism and suggestion form and then entering it in a suggestion box which is distinguished between satisfied and dissatisfied. However, such a method is less effective because it can be easily manipulated by customers and customers are often indifferent to the feedback. So that the improvement of service quality tends to be less effective. This research will design a system that can recognize human facial expressions to determine the level of customer satisfaction with input data in the form of video data taken by a webcam camera with the viola-jones method to detect faces and determine facial patterns. Then the facial data will be classified using the K-Nearest Neighbor method to determine the type of facial expression. Determination of the value of k will determine the success rate of facial expression detection. The processed data will be displayed on a liquid crystal display (LCD) and then stored in a MySQL database. The results showed that the accuracy of facial expression detection was 80.77% from 52 facial expression data.

Abstract WMP120: Development Of Smartphone Enabled Machine Learning Algorithms For Autonomous Stroke Detection

Background: We developed an automated smart phone application for detection of acute stroke using machine learning (ML) algorithms for recognition of facial asymmetry, arm weakness, and speech changes. Methods: We analysed prospectively collected data from patients admitted to 4 major metropolitan stroke centers with confirmed diagnosis of acute stroke. Speech and facial data were captured via video recording and arm data was captured via device sensors. A. Face. This module extracts 68 facial landmark points that are passed through a dimensionality reduction step and an asymmetry classifier. We implemented and compared 26 classification methods with neurologists' clinical impression and determined Quadratic Discriminative Analysis as the best one in terms of accuracy and interpretability. B. Arm. Using data extracted from 3D accelerometer, gyroscope, and magnetometer , we designed a grasp agnostic classifier based on AdaBoost to process motion trajectories and detect arm weakness.C. Speech. We developed an algorithm based on frequency analysis and Mel Frequency Cepstral Coefficients (MFCC) to detect abnormal/slurred speech. All tests were conducted within 72 hours of admission. Each of the three ML outputs was correlated with neurologists’ clinical impression. Results: Among the 269 analysed patients, 41% were female, the median age was 71, % had hemorrhagic and % had ischemic stroke. Final analyses of 18311 facial images revealed 99.42% sensitivity, 93.67% specificity, and 97.11% accuracy in detection of facial asymmetry. The results for 43 arm trajectories revealed 71.42% sensitivity, 72.41% specificity, and 72.09% accuracy in detection of arm weakness. Preliminary analysis of MFCC algorithms confirmed adequate features for abnormal speech detection Conclusions: Our preliminary results confirm that smartphone enabled ML-algorithms can reliably identify acute stroke features with accuracy comparable to neurologists’ clinical impression.

Real-world human gender classification from oral region using convolutional neural netwrok

Gender classification is an important biometric task. It has been widely studied in the literature. Face modality is the most studied aspect of human-gender classification. Moreover, the task has also been investigated in terms of different face components such as irises, ears, and the periocular region. In this paper, we aim to investigate gender classification based on the oral region. In the proposed approach, we adopt a convolutional neural network. For experimentation, we extracted the region of interest using the RetinaFace algorithm from the FFHQ faces dataset. We achieved acceptable results, surpassing those that use the mouth as a modality or facial sub-region in geometric approaches. The obtained results also proclaim the importance of the oral region as a facial part lost in the Covid-19 context when people wear facial mask. We suppose that the adaptation of existing facial data analysis solutions from the whole face is indispensable to keep-up their robustness.

Development of a Chinese werewolf deception database.

Although it is important to accurately detect deception, limited research in this area has been undertaken involving Asian people. We aim to address this gap by undertaking research regarding the identification of deception in Asians in realistic environments. In this study, we develop a Chinese Werewolf Deception Database (C2W2D), which consists of 168 video clips (84 deception videos and 84 honest videos). A total of 1,738,760 frames of facial data are recorded. Fifty-eight healthy undergraduates (24 men and 34 women) and 26 drug addicts (26 men) participated in a werewolf game. The development of C2W2D is accomplished based on a "werewolf" deception game paradigm in which the participants spontaneously tell the truth or a lie. Two synced high-speed cameras are used to capture the game process. To explore the differences between lying and truth-telling in the database, descriptive statistics (e.g., duration and quantity) and hypothesis tests are conducted using action units (AUs) of facial expressions (e.g., t-test). The C2W2D contributes to a relatively sizable number of deceptive and honest samples with high ecological validity. These samples can be used to study the individual differences and the underlying mechanisms of lies and truth-telling between drug addicts and healthy people.

Smart Facial Emotion Recognition With Gender and Age Factor Estimation

Human-Computer Interaction (HCI) in an intelligent way, which aims at creating scalable and flexible solutions. Big tech firms and businesses believe in the success of HCI as it allows them to profit from on-demand technology and infrastructure for information-centric applications without having to use public clouds. Because of its capacity to imitate human coding abilities, facial expression recognition and software-based facial expression identification systems are crucial. This paper proposes a system of recognizing the emotional condition of humans, given a facial expression, and conveys two methods of predicting the age and gender factors from human faces. This research also aims in understanding the influences posed by gender and age of humans on their facial expressions. The model can currently detect 7 emotions based on the facial data of a person - (Anger, Disgust, Happy, Fear, Sad, Surprise, and Neutral state). The proposed system is divided into three segments: a.) Gender Detection b.) Age Detection c.) Emotion Recognition. The initial model is created using 2 algorithms - KNN, and SVM. We have also utilized the architectures of some of the deep learning models such as CNN and VGG - 16 pre-trained models (Transfer Learning). The evaluation metrics show the model performance regarding the accuracy of the Recognition system. Future enhancements of this work can include the deployment of the DL and ML model onto an android or a wearable device such as a smartphone or a watch for a real-time use case.

A Convolutional Neural Network Face Recognition Method Based on BiLSTM and Attention Mechanism.

Face recognition technology is a powerful means to capture biological facial features and match facial data in existing databases. With the advantages of noncontact and long-distance implementation, it is being used in more and more scenarios. Affected by factors such as light, posture, and background environment, the face images captured by the device are still insufficient in the recognition rate of existing face recognition models. We propose an AB-FR model, a convolutional neural network face recognition method based on BiLSTM and attention mechanism. By adding an attention mechanism to the CNN model structure, the information from different channels is integrated to enhance the robustness of the network, thereby enhancing the extraction of facial features. Then, the BiLSTM method is used to extract the timing characteristics of different angles or different time photos of the same person so that convolutional blocks can obtain more face detail information. Finally, we used the cross-entropy loss function to optimize the model and realize the correct face recognition. The experimental results show that the improved network model indicates better identification performance and stronger robustness on some public datasets (such as CASIA-FaceV5, LFW, MTFL, CNBC, and ORL). Besides, the accuracy rate is 99.35%, 96.46%, 97.04%, 97.19%, and 96.79%, respectively.

A Systematic Approach to Adaptive Mesh Refinement for Computational Electrodynamics

There is a great need to solve CED problems on adaptive meshes; referred to here as AMR-CED. The problem was deemed to be susceptible to “long-term instability” and parameterized methods have been used to control the instability. In this paper, we present a new class of AMR-CED methods that are free of this instability because they are based on a more careful understanding of the constraints in Maxwell's equations and their preservation on a single control volume. The important building blocks of these new methods are: 1) Timestep sub-cycling of finer child meshes relative to parent meshes. 2) Restriction of fine mesh facial data to coarser meshes when the two meshes are synchronized in time. 3) Divergence constraint-preserving prolongation of the coarse mesh solution to newly built fine meshes or to the ghost zones of pre-existing fine meshes. 4) Electric and magnetic field intensity-correction strategy at fine-coarse interfaces. Using examples, we show that the resulting AMR-CED algorithm is free of “long-term instability”. Unlike previous methods, there are no adjustable parameters. The method is inherently stable because a strict algorithmic consistency is applied at all levels in the AMR mesh hierarchy. We also show that the method preserves order of accuracy, so that high order methods for AMR-CED are indeed possible.