Frontal View Face Research Articles

Facial Segmental Infantile Haemangioma: A Cutaneous Mosaicism Quandary.

Mosaicism has long been considered the underlying mechanism of segmental infantile hemangiomas (SIH). This was a prospective pilot case-control study conducted with the objective to quantify the percentage overlap of silhouettes of facial SIH with those of Blaschko lines (the most well studied archetypical pattern of mosaicism on face) as compared to other mosaic disorders on face. Lesional silhouettes of 8 patients with SIH (Group A) and 6 patients with other facial dermatosis known to have blaschkoidal distribution (Group B), were overlapped on a standardized template with Blaschkoidal lines on the frontal view of face. The alignment was done via the auto align tool of Photoshop and the percentage of overlap was calculated with an online image comparison software (IMGonline.com.ua). There was a significant difference in mean overlap in Group A (72.92 ± 15.6 %) as compared to Group B (90.1 ± 4.3%; P=0.018). Hence, we concluded that facial SIH do not follow lines of Blaschko.

Generative Adversarial Network Based Face Frontalization: A Review

Abstract: Face frontalization can boost the performance of face recognition methods and has made significant progress with the development of Generative Adversarial Networks (GANs). Face frontalization, the process of synthesizing frontal views of faces from arbitrary poses, plays a pivotal role in numerous computer vision applications including facial recognition, emotion analysis, and virtual reality. In this study, we propose a novel approach leveraging Generative Adversarial Networks (GANs) to address the challenging task of face frontalization. Our model integrates a carefully designed architecture capable of disentangling pose variations from facial features. Additionally, a multi-stage training strategy is employed to enhance the network's capacity for learning complex pose-to-pose mappings

Understanding face detection with visual arrays and real-world scenes

ABSTRACT Face detection has been studied by presenting faces in blank displays, object arrays, and real-world scenes. This study investigated whether these display contexts differ in what they can reveal about detection, by comparing frontal-view faces with those shown in profile (Experiment 1), rotated by 90° (Experiment 2), or turned upside-down (Experiment 3). In blank displays, performance for all face conditions was equivalent, whereas upright frontal faces showed a consistent detection advantage in arrays and scenes. Experiment 4 examined which facial characteristics drive this detection advantage by rotating either the internal or external facial features by 90° while the other features remained upright. Faces with rotated internal features were detected as efficiently as their intact frontal counterparts, whereas detection was impaired when external features were rotated. Finally, Experiment 5 applied Voronoi transformations to scenes to confirm that complexity of stimulus displays modulates the detection advantage for upright faces. These experiments demonstrate that context influences what can be learned about the face detection process. In complex visual arrays and natural scenes, detection proceeds more effectively when external facial features are preserved in an upright orientation. These findings are consistent with a cognitive detection template that focuses on general face-shape information.

Things are looking up: Vertical eye gaze in the environment affects perceptions of emotional valence in sad faces.

Successful social interactions depend on the ability to quickly evaluate emotional facial expressions. Research has shown that head orientation and eye gaze are informative affective signals. Across four experiments, we explored a novel eye-gaze cue grounded in a consideration of English spatial metaphors, where up connotes positive feelings ("I'm flying high") and down connotes negative feelings ("I'm feeling low"). Participants either rated the valence of or categorised a set of sad and happy faces gazing in different directions along the vertical axis. We expected to find a spatial-valence congruency effect, where valence ratings and reaction times would be moderated by whether or not the face was gazing in a metaphor-consistent direction. The results partially supported this hypothesis: sad faces gazing upwards (as opposed to downwards) were rated as happier or more positive (Experiments 1 and 2) and classified slower (Experiments 3 and 4). This was true whether the looking direction was cued by eye gaze in front-view faces (Experiment 1) or by the orientation of profile faces (Experiments 2-4). In addition, this spatial-valence congruency effect was only reliable in the environmental frame of reference (Experiment 4). We found little evidence for a comparable effect of gaze direction on judgements of happy faces, suggesting that eye gaze along the vertical axis may differentially affect judgements of approach and avoidance-related emotional expressions. This has implications for the inferences scholars draw about underlying cognitive representations from observations of conventional metaphorical language.

Multimodal soft biometrics: combining ear and face biometrics for age and gender classification

In this paper, we present a multimodal, multitask deep convolutional neural network framework for age and gender classification. In the developed framework, we have employed two different biometric modalities: ear and profile face. We have explored three different fusion methods, namely, data, feature, and score fusion, to combine the information extracted from ear and profile face images. In the framework, we have utilized VGG-16 and ResNet-50 models with center loss to obtain more discriminative features. Moreover, we have performed two-stage fine-tuning to increase the representation capacity of the models. To assess the performance of the proposed approach, we have conducted extensive experiments on the FERET, UND-F, and UND-J2 datasets. Experimental results indicate that ear and profile face images contain useful features to extract soft biometric traits. We have shown that when frontal face view of the subject is not available, use of ear and profile face images can be a good alternative for the soft biometric recognition systems. The presented multimodal system achieves very high age and gender classification accuracies, matching the ones obtained by using frontal face images. The multimodal approach has outperformed both the unimodal approaches and the previous state-of-the-art profile face image or ear image-based age and gender classification methods, significantly in both tasks.

Face Detection of Thermal Images in Various Standing Body-Pose using Facial Geometry

Automatic face detection in frontal view for thermal images is a primary task in a health system e.g. febrile identification or security system e.g. intruder recognition. In a daily state, the scanned person does not always stay in frontal face view. This paper develops an algorithm to identify a frontal face in various standing body-pose. The algorithm used an image processing method where first it segmented face based on human skin’s temperature. Some exposed non-face body parts could also get included in the segmentation result, hence discriminant features of a face were applied. The shape features were based on the characteristic of a frontal face, which are: (1) Size of a face, (2) facial Golden Ratio, and (3) Shape of a face is oval. The algorithm was tested on various standing body-pose that rotate 360° towards 2 meters and 4 meters camera-to-object distance. The accuracy of the algorithm on face detection in a manageable environment is 95.8%. It detected face whether the person was wearing glasses or not.

Semi-Supervised Face Frontalization in the Wild

Synthesizing a frontal view face from a single nonfrontal image, i.e. face frontalization, is a task of practical importance in a wide range of facial image analysis applications. However, to train the frontalization model in a supervised manner, most existing face frontalization methods rely on the availability of nonfrontal-frontal face pairs (typically from the Multi-PIE dataset) captured in a constrained environment. Such approaches, in return, limit the generalizability of their application to unconstrained scenarios. Unfortunately, although a large amount of in-the-wild face datasets are available, they cannot easily be utilized for face frontalization training since the nonfrontal and frontal facial images are not paired. To train a frontalization network which generalizes well to both constrained and unconstrained environments, we propose a semi-supervised learning framework which effectively uses both (labeled) indoor and (unlabeled) outdoor faces. Specifically, to achieve this goal, this article presents a Cycle-Consistent Face Frontalization Generative Adversarial Network (CCFF-GAN) which consists of both (1) the supervised and (2) the unsupervised components. For (1), we use the indoor paired (labeled) data to learn a roughly accurate frontalization network which may not generalize well to outdoor (in-the-wild) scenarios. For (2), to cope with the generalization issue, the unsupervised part uses the unpaired (unlabeled) images under the perceptual cycle consistency constraint in the semantic feature space to generalize the network from controlled (indoor) to uncontrolled (outdoor) environment. Extensive experiments demonstrate the effectiveness of the proposed method in comparison with the state-of-the-art face frontalization methods, especially under the in-the-wild scenarios.

2D Pose-Invariant Face Recognition Using Single Frontal-View Face Database

Personal identification systems that use face recognition work well for test images with frontal view face, but often fail when the input face is a pose view. Most face databases come from picture ID sources such as passports or driver’s licenses. In such databases, only the frontal view is available. This paper proposes a method of 2D pose-invariant face recognition that assumes the search database contains only frontal view faces. Given a non-frontal view of a test face, the pose-view angle is first calculated by matching the test image with a database of canonical faces with head rotations to find the best matched image. This database of canonical faces is used only to find the head rotation. The database does not contain images of the test face itself, but has a selection of template faces, each face having rotation images of − 45°, − 30°, − 15°, 0°, 15°, 30°, and 45°. The landmark features in the best matched rotated canonical face such as say rotation 15° and it’s corresponding frontal face of rotation 0° are used to create a warp transformation to convert the 15° rotated test face to a frontal face. This warp will introduce some distortion artifacts since some features of the non-frontal input face are not visible due to self-occlusion. The warped image is, therefore, enhanced by mixing intensities using the left/right facial symmetry assumption. The enhanced synthesized frontal face image is then used to find the best match target in the frontal face database. We test our approach using CMU Multi-PIE database images. Our method performs with acceptable and similar accuracy to conventional methods, while using only frontal faces in the test database.

Parametric face drawings: A demographically diverse and customizable face space model.

We introduce a novel face space model-parametric face drawings (or PFDs)-to generate schematic, though realistic, parameterized line drawings of faces based on the statistical distribution of human facial features. A review of existing face space models (including FaceGen Modeller, Synthetic Faces, MPI, and active appearance model) indicates that current models are constrained by their reliance on ethnically homogeneous face databases. This constraint has led to negative consequences for underrepresented populations, such as impairments in automatized identity recognition of certain demographic groups. Our model is based on a demographically diverse sample of 400 faces (200 female, 200 male; 100 East Asian/Pacific Islander, 100 Latinx/Hispanic, 100 black/African-American, and 100 white/Caucasian) compiled from several face databases (including FERET face recognition technology and the Chicago Face Database). Each front-view face image is manually coded with 85 landmark points that are then normalized and rendered with MATLAB (MathWorks, Natick, MA) tools to produce a smooth, parameterized face line drawing. We present data from two behavioral experiments to validate our model and demonstrate its applicability. In Experiment 1 we show that PFDs produce a reliable "inversion effect" in short-term recognition, a hallmark of holistic processing. In Experiment 2, we conduct a celebrity recognition task, comparing performance on PFDs to performance on untextured renderings from FaceGen Modeller. Participants successfully recognized approximately 50% of celebrity faces based on the PFD models, comparable to performance based on FaceGen Modeler (also 50% correct). We highlight a range of potential applications of our model, list some limitations, and provide MATLAB resources for researchers to utilize our face space, including the ability to customize the demographic makeup of the face space, add new faces, and produce morphs and caricatures.

An improved SIFT algorithm for robust emotion recognition under various face poses and illuminations

To address the variabilities of the number and position of extracted feature points for the traditional scale-invariant feature transform (SIFT) method, an improved SIFT algorithm is proposed for robust emotion recognition. Specifically, shape decomposition is first performed on the detected facial images by defining a weight vector. Then, a feature point constraint algorithm is developed to determine the optimum position of the feature points that can effectively represent the expression change regions. On this basis, the SIFT descriptors are applied to extract the regional gradient information as feature parameters. Finally, the support vector machine classifier combined with the principal component analysis method is used to reduce the feature dimensions and facial expression recognition. Experiments have been performed under different conditions, i.e., varied illuminations, face poses and facial moisture levels, using 15 participants. In the cases of frontal face and 5-degree face rotation views, the average recognition accuracies are 98.52% and 94.47% (no additional light sources), as well as 96.97% and 95.40% (two additional light sources), respectively. In addition, as an effective supplement to the problem of changes in illumination, the average recognition ratios are 96.23% and 96.20% under dry and wet face conditions, respectively. The experimental results reveal the robust performance of the proposed method in facial expression recognition.

A Validated Assessment Scale for Asian Chin Projection.

As the number of different aesthetic treatments increase, numerous photonumeric assessment scales have been developed and validated to measure the effectiveness of these new treatments and techniques. Photonumeric rating scales have been developed to objectively assess improvements in anatomical areas; however, these have been based on the features of Caucasian patients. To develop and validate a Chin Projection Scale for use in the female Asian patient population. During 2 validation sessions, 13 raters assessed full frontal and lateral facial views of 50 Asian subjects and also estimated their age and the aesthetic treatment effort required for each subject. Chin projection was rated on a scale from 0 (optimal) to 4 (very severely receding). Inter-rater reliability was 0.80 (substantial) for Validation Session 1 and 0.83 (almost perfect) for Validation Session 2. The results for Estimated Age and Estimated Treatment Effort were essentially the same. This study demonstrated the validity of the first photonumeric assessment scale for assessing the appearance of the female Asian chin. This new scale will provide a standardized measure of chin projection for Asian patients in clinical practice and clinical research settings.

Object recognition in fish: accurate discrimination across novel views of an unfamiliar object category (human faces)

Accurate visual object recognition is essential to survival for a wide range of species across a wide range of evolutionary histories and visual requirements. However this task is solved, it is a major achievement because object recognition is far from simple. The appearance of an object can alter almost completely as viewing conditions change, not least under variations in lighting and orientation. Determining the recognition limits of a species is important to understanding its visual ecology and can help identify conditions under which recognition may fail. In this study, we tested whether a species of fish can recognize objects from an unfamiliar object class (human faces) across changes in viewing direction. Using operant conditioning, we trained archerfish, Toxotes chatareus, to discriminate between two frontal views of standardized human faces and, critically, tested whether they could continue to do so as the orientation in depth of the faces changed. All fish learned the initial discrimination task and could also recognize rotated forms. These results represent the first conclusive evidence that a species of fish can generalize recognition across views, speaking against a strict image-matching process. This ability rather speaks to the capacity of relatively simple brains to tackle the hard problem of view invariance and provides insight into the mechanisms employed in more complex organisms such as humans. Although we speculate that other fish species may demonstrate similar abilities, a visual system capable of recognition across changes in viewpoint may be especially important to the unique hunting strategy of archerfish.

Temporal Contiguity Training Influences Behavioral and Neural Measures of Viewpoint Tolerance.

Humans can often recognize faces across viewpoints despite the large changes in low-level image properties a shift in viewpoint introduces. We present a behavioral and an fMRI adaptation experiment to investigate whether this viewpoint tolerance is reflected in the neural visual system and whether it can be manipulated through training. Participants saw training sequences of face images creating the appearance of a rotating head. Half of the sequences showed faces undergoing veridical changes in appearance across the rotation (non-morph condition). The other half were non-veridical: during rotation, the face simultaneously morphed into another face. This procedure should successfully associate frontal face views with side views of the same or a different identity, and, according to the temporal contiguity hypothesis, thus enhance viewpoint tolerance in the non-morph condition and/or break tolerance in the morph condition. Performance on the same/different task in the behavioral experiment (N = 20) was affected by training. There was a significant interaction between training (associated/not associated) and identity (same/different), mostly reflecting a higher confusion of different identities when they were associated during training. In the fMRI study (N = 20), fMRI adaptation effects were found for same-viewpoint images of untrained faces, but no adaptation for untrained faces was present across viewpoints. Only trained faces which were not morphed during training elicited a slight adaptation across viewpoints in face-selective regions. However, both in the behavioral and in the neural data the effects were small and weak from a statistical point of view. Overall, we conclude that the findings are not inconsistent with the proposal that temporal contiguity can influence viewpoint tolerance, with more evidence for tolerance when faces are not morphed during training.

Face analysis in video: face detection and tracking with pose estimation

We introduced a full automatic approach to achieve face detection and tracking with pose estimation in video sequences. The proposed approach consists of three modules: face detection module, face tracking module and face pose estimation module. A combination between detection and tracking modules was performed to overcome the different challenging problems that might occur while detecting or tracking faces. Afterward face pose estimation module was applied to select the best camera capture which is closest to the frontal face view for better face recognition task. The performance of these modules was evaluated with an experimental study which has proven the robustness of the proposed approach for a face analysis in video.

Differences between Caucasian and Asian attractive faces

There are discrepancies between the public's current beauty desires and conventional theories and historical rules regarding facial beauty. This photogrammetric study aims to describe in detail mathematical differences in facial configuration between attractive Caucasian and attractive Asian faces. To analyse the structural differences between attractive Caucasian and attractive Asian faces, frontal face and lateral face views for each race were morphed; facial landmarks were defined, and the relative photographic pixel distances and angles were measured. Absolute values were acquired by arithmetic conversion for comparison. The data indicate that some conventional beliefs of facial attractiveness can be applied but others are no longer valid in explaining perspectives of beauty between Caucasians and Asians. Racial differences in the perceptions of attractive faces were evident. Common features as a phenomenon of global fusion in the perspectives on facial beauty were revealed. Beauty standards differ with race and ethnicity, and some conventional rules for ideal facial attractiveness were found to be inappropriate. We must reexamine old principles of facial beauty and continue to fundamentally question it according to its racial, cultural, and neuropsychological aspects.

Automated face detection for occurrence and occupancy estimation in chimpanzees.

Using semi-automated ape face detection technology for processing camera trap footage requires only 2-4% of the time compared to manual analysis and allows to estimate site use by chimpanzees relatively reliably.

Perception of Acceptable Range of Smiles by Specialists, General Dentists and Lay Persons and Evaluation of Different Aesthetic Paradigms.

One of the most important goals of restorative dentistry is to restore the patient's aesthetic. Smile analysis is subjective and it differs from person to person. An aesthetic smile involves a harmonious relationship between various parameters including the hard and soft tissues. The aim of the study was to identify the acceptable range of several smiles (alone and in conjunction with the face) by specialists, general dentists as well as lay persons; and to identify the values of different criteria i.e., the Golden Proportion (GP), the Recurrent Esthetic Dental proportion (RED), Width to Height ratio (W/H ratio), the Apparent Contact Dimension (ACD), and lateral incisor position in a smile. Hundred photographs of 50 subjects were taken, 50 of the smile alone and 50 of the individual's frontal view of face. The photographs of the smiles and the faces were assessed for the aesthetic acceptability by 30 evaluators including 10 specialists with advanced training, 10 general dentists and 10 lay persons. Irreversible hydrocolloid impressions were made of the dentitions of all the individuals using stock trays and were poured in dental stone. Measurements were made on the facial surface of the teeth on the models and were recorded in millimeters using a sharp tipped digital vernier calliper. Data was analyzed to evaluate the presence of different parameters assessed in the smiles. Mean and standard deviation values for the percentage of only the agreeable smiles were calculated in both individual smile analysis and in conjunction with the face. The non agreeable smiles were excluded from further statistical analysis. Pearson Correlation Coefficient was calculated to compare the values obtained in all the three groups. More number of smiles were considered agreeable by the general dentists when compared to the specialists and the number even increased in case of evaluation by lay persons. Greater number of smiles was found to be agreeable when they were evaluated in conjunction with the face. Rather than assessment of individual numeric parameter that defines an ideal aesthetic smile, a smile to be aesthetic should harmonize with the composition of the face.

An Analysis of Adaboost Algorithm for Face Detection

Objective: The goal of this paper is to analyze the computation of detecting faces from still images using adaboost algorithm. Methods: The dataset is collected from the web. The chosen image is converted into gray scale image. Sum of the pixel values are calculated with the utilization of summed area table. For each step of iteration, adaboost learn and reconstruct the weight value. For classification, cascade classifiers are used to differentiate face and non-face regions. Findings: The method is suitable for frontal view face images only.

Efficient visual information for unfamiliar face matching despite viewpoint variations: It’s not in the eyes!

Faces are encountered in highly diverse angles in real-world settings. Despite this considerable diversity, most individuals are able to easily recognize familiar faces. The vast majority of studies in the field of face recognition have nonetheless focused almost exclusively on frontal views of faces. Indeed, a number of authors have investigated the diagnostic facial features for the recognition of frontal views of faces previously encoded in this same view. However, the nature of the information useful for identity matching when the encoded face and test face differ in viewing angle remains mostly unexplored. The present study addresses this issue using individual differences and bubbles, a method that pinpoints the facial features effectively used in a visual categorization task. Our results indicate that the use of features located in the center of the face, the lower left portion of the nose area and the center of the mouth, are significantly associated with individual efficiency to generalize a face’s identity across different viewpoints. However, as faces become more familiar, the reliance on this area decreases, while the diagnosticity of the eye region increases. This suggests that a certain distinction can be made between the visual mechanisms subtending viewpoint invariance and face recognition in the case of unfamiliar face identification. Our results further support the idea that the eye area may only come into play when the face stimulus is particularly familiar to the observer.

Review and testing of frontal face detectors

This paper presents comparison results for the proposed face detection algorithm based on a compact convolutional neural network cascade and modern frontal face detectors. Test results for 16 frontal view face detectors on two public benchmarks datasets are shown. A comparative assessment of the performance of face detection algorithms is made.