Full-face Images Research Articles

PrivateGaze: Preserving User Privacy in Black-box Mobile Gaze Tracking Services

Eye gaze contains rich information about human attention and cognitive processes. This capability makes the underlying technology, known as gaze tracking, a critical enabler for many ubiquitous applications and has triggered the development of easy-to-use gaze estimation services. Indeed, by utilizing the ubiquitous cameras on tablets and smartphones, users can readily access many gaze estimation services. In using these services, users must provide their full-face images to the gaze estimator, which is often a black box. This poses significant privacy threats to the users, especially when a malicious service provider gathers a large collection of face images to classify sensitive user attributes. In this work, we present PrivateGaze, the first approach that can effectively preserve users' privacy in black-box gaze tracking services without compromising gaze estimation performance. Specifically, we proposed a novel framework to train a privacy preserver that converts full-face images into obfuscated counterparts, which are effective for gaze estimation while containing no privacy information. Evaluation on four datasets shows that the obfuscated image can protect users' private information, such as identity and gender, against unauthorized attribute classification. Meanwhile, when used directly by the black-box gaze estimator as inputs, the obfuscated images lead to comparable tracking performance to the conventional, unprotected full-face images.

High-Fidelity Synthetic Face Generation for Rosacea Skin Condition from Limited Data

Similarly to the majority of deep learning applications, diagnosing skin diseases using computer vision and deep learning often requires a large volume of data. However, obtaining sufficient data for particular types of facial skin conditions can be difficult, due to privacy concerns. As a result, conditions like rosacea are often understudied in computer-aided diagnosis. The limited availability of data for facial skin conditions has led to the investigation of alternative methods of computer-aided diagnosis. In recent years, generative adversarial networks (GANs), mainly variants of StyleGANs, have demonstrated promising results in generating synthetic facial images. In this study, for the first time, a small dataset of rosacea with 300 full-face images was utilized to further investigate the possibility of generating synthetic data. Our experimentation demonstrated that the strength of R1 regularization is crucial for generating high-fidelity rosacea images using a few hundred images. This was complemented by various experimental settings to ensure model convergence. We successfully generated 300 high-quality synthetic images, significantly contributing to the limited pool of rosacea images for computer-aided diagnosis. Additionally, our qualitative evaluations by 3 expert dermatologists and 23 non-specialists highlighted the realistic portrayal of rosacea features in the synthetic images. We also provide a critical analysis of the quantitative evaluations and discuss the limitations of solely relying on validation metrics in the field of computer-aided clinical image diagnosis.

Comparison of the Identification of Emotional Facial Expressions from Full Faces with only the Eyes and the Mouth Regions in Young Indian Adults: an Exploratory Study.

Identification of emotional facial expressions (EFEs) is important in interpersonal communication. Six 'universal' EFEs are known, though accuracy of their identification varies. EFEs involve anatomical changes in certain regions of the face, especially eyes and mouth. But whether other areas of the face are just as important in their identification is still debated. This study was conducted to compare the accuracy of identification of universal EFEs under full-face and partial face conditions (only showing the eyes and the mouth regions). An analytical cross-sectional study was conducted among 140 young Indian adults. They were divided into two equal groups and shown the six universal EFEs in two sets, one with full-face images, and the other with images showing just the eyes and the mouth regions on a computer screen. The participants were asked to identify each of the EFE and their responses were analyzed. Mean age was 21.3 1.7 years for full face group, and 21.2 1.6 years for the partial face group. Most were men, from rural areas and from upper socioeconomic status families, and many of them were students. EFE identification was significantly higher for part-face group compared to full-face group (p-value .0007). Participants of both groups identified happiness the best (100%). For other EFEs, part-face images were identified more accurately than full-face images, except for disgust. These differences were statistically significant except for anger and fear. Among young Indian adults, accuracy of identification of universal EFEs was high, which was significantly enhanced for all except disgust, when only combinations of eyes and mouth were shown, suggesting that other facial regions serve as distractors in EFE identification. 1. Identification of universal EFEs was higher from partial faces (combination of eyes and mouth) as compared to full-face EFEs for all emotions except disgust. 2. This suggests that other regions of face serve as potential distractors in the identification of emotions, except for disgust, where these regions provide more information.

Role of Dentofacial Harmony among Young and Old Adult Females in Smile Perception of Dental Specialists and Laypeople: A Cross-Sectional Study.

The purpose of this study was to determine whether balance and harmony of the face, when seen as a whole, alter the smile perception as compared to a lone standing smile in female subjects of two age groups and to identify parameters of dental composition and dentofacial harmony associated with the smile perception of dental specialists (DS) and laypeople (LP). A cross-sectional study. Twenty-nine DS and 29 LP scored the smile attractiveness of 84 full-face images (FFIs) and corresponding smile images (SIs) of young adults (YA: 20-29 years) and old adults (OA: 50-59 years), using the Visual Analogue Scale. Smile analysis software (Planmeca Romexis®Smile-Design) was used to measure 10 parameters of dental composition and dentofacial harmony in FFI. Independent t-test and Pearson's correlation coefficient (r) were used to analyze the data. A significant difference was not seen between the smile attractiveness scores of FFI and SI in both raters and age groups. Significant differences were observed between DS and LP in the YA group. The OA group was rated significantly higher than the YA group. The correlation between smile attractiveness scores and each of the nine dentofacial parameters was not significant (P < 0.05). Framing of images did not significantly alter smile perception. FFI received higher scores than SI, and raters were more critical of the YA group. Professional training of raters and an interplay of smile and facial attractiveness of the female subjects contributed to the smile perception of raters. No parameters of dental composition or dentofacial harmony significantly associated with smile attractiveness were identified.

Person-Specific Gaze Estimation from Low-Quality Webcam Images.

Gaze estimation is an established research problem in computer vision. It has various applications in real life, from human-computer interactions to health care and virtual reality, making it more viable for the research community. Due to the significant success of deep learning techniques in other computer vision tasks-for example, image classification, object detection, object segmentation, and object tracking-deep learning-based gaze estimation has also received more attention in recent years. This paper uses a convolutional neural network (CNN) for person-specific gaze estimation. The person-specific gaze estimation utilizes a single model trained for one individual user, contrary to the commonly-used generalized models trained on multiple people's data. We utilized only low-quality images directly collected from a standard desktop webcam, so our method can be applied to any computer system equipped with such a camera without additional hardware requirements. First, we used the web camera to collect a dataset of face and eye images. Then, we tested different combinations of CNN parameters, including the learning and dropout rates. Our findings show that building a person-specific eye-tracking model produces better results with a selection of good hyperparameters when compared to universal models that are trained on multiple users' data. In particular, we achieved the best results for the left eye with 38.20 MAE (Mean Absolute Error) in pixels, the right eye with 36.01 MAE, both eyes combined with 51.18 MAE, and the whole face with 30.09 MAE, which is equivalent to approximately 1.45 degrees for the left eye, 1.37 degrees for the right eye, 1.98 degrees for both eyes combined, and 1.14 degrees for full-face images.

Healthy aging impairs face discrimination ability.

Face images enable individual identities to be discriminated from one another. We aimed to quantify age-related changes in different aspects of face identity discrimination. Face discrimination sensitivity was measured with a memory-free “odd-one-out” task. Five age groups (N = 15) of healthy adults with normal vision were tested: 20, 50-59, 60-69, 70-79, and 80-89. Sensitivity was measured for full-face images (all features visible), external features (head-shape, hairline), internal features (nose, mouth, eyes, and eyebrows) and closed-contour shapes (control object). Sensitivity to full-faces continuously declined by approximately 13% per decade, after 50 years of age. When age-related differences in visual acuity were controlled, the effect of age on face discrimination sensitivity remained. Sensitivity to face features also deteriorated with age. Although the effect for external features was similar to full-faces, the rate of decline was considerably steeper (approximately 3.7 times) for internal, relative to external, features. In contrast, there was no effect of age on sensitivity to shapes. All age groups demonstrated the same overall pattern of sensitivity to different types of face information. Healthy aging was associated with a continuous decline in sensitivity to both full-faces and face features, although encoding of internal features was disproportionately impaired. This age-related deficit was independent of differences in low-level vision. That sensitivity to shapes was unaffected by age suggests these results cannot be explained by general cognitive decline or lower-level visual deficits. Instead, healthy aging is associated with a specific decline in the mechanisms that underlie face discrimination.

Relationship of Various Anatomic Landmarks with the Midline of Face and Oral Commissure

Introduction: Currently, there are no exact guidelines which to choose specific anatomic landmarks while determining the midline of the face or midline of the mouth. The purpose of this study was to determine the hierarchy of facial anatomic landmarks closest to the midline of the face as well as midline of the oral commissure mouth. Methods: Anatomical landmarks Nasion, tip of the nose, and tip of the philtrum, were marked clinically on 100 subjects (age range: 20-35 years). Frontal full-face digital images of the subjects in smiling position were then made under standardized conditions. A total of 97 subjects met the inclusion criteria. Upon applying exclusion criteria, images of 73 subjects were used for midline analysis using a concept called the Esthetic Frame. One sample t tests were conducted at alpha values of .005.Results: The results indicated that anatomic landmarks soft tissue nasion, tip of nose, tip of papillae and dental midline deviated uniquely and significantly (P &lt;.005) from the midlines of the face as well as the mouth (oral commissure). Conclusions: Within the limitations of the study, the hierarchy of anatomic landmarks closest to the midline of the face in smile was as follows: the midline of the oral commissures, natural dental midline, tip of philtrum, nasion, and tip of the nose. The hierarchy of anatomic landmarks closest to the midline of the oral commissures was: natural dental midline, tip of philtrum, nasion and tip of nose.

Photographic Assessment of Intermittent Exotropia Using a Translucent Cover.

To evaluate the validity of the photographic assessment with a translucent cover that is clinically available for the evaluation of intermittent exotropia. Photographs of 270 patients who cooperated for the prism and alternate cover test (PACT) were reviewed. Full-face images were obtained with a digital camera while a translucent cover was placed in front of either eye. The change in distance from the medial canthus to the nasal limbus with occlusion was measured and the photographic angle was estimated by three independent ophthalmologists based on representative photographs. These two measurements were correlated with the angle measured using the PACT, and clinical features related to the discrepancy between the photographic angle and the angle measured with the PACT were determined. Patients with intermittent exotropia of 27.0 ± 6.1 prism diopters (PD) showed a 4.5 ± 3.3 mm exodrift and an estimated angle of 29.0 ± 4.3 PD on the photographs with occlusion. The exodrift distance and photographic angle were positively correlated with the angle of PACT (r = 0.256, P < .001 and r = 0.546, P < .001, respectively). Of the 47 patients with a discrepancy of greater than 8 PD, 38 patients (80.9%) were regarded as having a photographic angle that was larger than the angle of PACT, which was more common in older patients, in those with a small distance angle, and when taking off the spectacles. Photographs with a translucent cover can reveal the latent components of intermittent exotropia. However, the photographic angle might differ from the real angle, particularly in older patients with a small angle and with the spectacles off. [J Pediatr Ophthalmol Strabismus. 2021;58(5):331-338.].

Effect of variations in facial flow curves on the perceptions of smile esthetics by laypeople

Statement of problemInformation about the effect of the facial flow concept on the smile esthetic perception of laypeople is lacking. PurposeThe purpose of this study was to determine the effect of the facial flow concept on laypeople’s perception of smile esthetics. Material and methodsFrontal full-face images of a female volunteer were digitally altered to create commissural line and transverse occlusal plane inclinations, a dental midline shift, and facial asymmetries. A questionnaire was developed, and a single researcher asked 400 evaluators to rate the attractiveness of the different smiles by using a visual analog scale. The subsequent data were analyzed by using 3-factor repeated measures ANOVA and post hoc Bonferroni tests. ResultsImages with commissural line and transverse occlusal plane inclinations pointing toward the green side of the facial flow curve were found to be more attractive than those with inclinations pointing to the red side (P<.001). Asymmetric facial images with the dental midline coinciding with the facial flow curve had higher esthetic scores than those without (P<.001). Images that showed parallelism between the transverse occlusal plane and commissural line inclinations were perceived as more esthetic (P<.001). ConclusionsIn an asymmetric face model, the degree and direction of commissural line and transverse occlusal plane inclinations and dental midline shift influenced the perceived attractiveness of a smile.

Half-face based recognition using principal component analysis

&lt;span lang="EN-US"&gt;Face recognition is a considerable problem in the field of image processing. It is used daily in various applications from personal cameras to forensic investigations. Most of the provides solutions proposed based on full-face images, are slow to compute and need more storage. In this paper, we propose an effective way to reduce the features and size of the database in the face recognition method and thus we get an increase in the speed of discrimination by using half of the face. Taking advantage of face symmetry, the first step is to divide the face image into two halves, then the left half is processed using the principal component analysis (PCA) algorithm, and the results are compared by using Euclidian distance to distinguish the person. The system was trained and tested on ORL database. It was found that the accuracy of the system reached up to 96%, and the database was minimized by 46% and the running time was decreased from 120 msec to 70 msec with a 41.6% reduction.&lt;/span&gt;

Inter-observer evaluation of erythema-directed photography for the assessment of erythema and telangiectasias in rosacea.

Persistent centrofacial erythema associated with telangiectasias is one of the most common phenotypes of rosacea in clinical practice, and the assessment of each component is crucial as each of them may require a different approach. The aim of this study was to evaluate the inter-observer reliability of standard photography vs erythema-directed photography for the assessment of erythema and telangiectasias in rosacea. One hundred full-face images of 50 rosacea patients (50 standard photographs and 50 erythema-directed digital photographs) were evaluated by 8 independent experienced dermatologists using a 5-item score for erythema and telangiectasias, respectively. Inter-rater reliability, by comparing erythema and telangiectasias scores and calculating the percentage of agreement between evaluators, was assessed and the strength of agreement using the Cohen's Kappa values (95% CI) was calculated. Poor and fair strength of agreement for erythema and telangiectasias using standard photography vs moderate and good strength of agreement using erythema-directed digital photography was found. Erythema-directed digital photography may provide a better strength of agreement and higher reliability among independent observers compared to standard photography in the assessment of erythema and telangiectasias in rosacea, thus suggesting new horizons for digital appraisal of skin diseases.

An Artificial Intelligence Approach to the Assessment of Abnormal Lid Position.

Summary:New artificial intelligence (AI) approaches to facial analysis show promise in the clinical evaluation of abnormal lid position. This could allow more naturalistic, quantitative, and automated assessment of lid position. The aim of this article was to determine whether OpenFace, an AI approach to real-time facial landmarking and analysis, can extract clinically useful measurements from images of patients before and after ptosis correction. Manual and AI-automated approaches to vertical palpebral aperture measurement of 128 eyes in pre- and postoperative full-face images of ptosis patients were compared in this study. Agreement in interpupillary distance to vertical palpebral aperture ratio between clinicians and an AI-based system was assessed. Image quality varied highly with interpupillary distance defined by a mean of 143.4 pixels (min = 60, max = 328, SD = 80.3 pixels). A Bland–Altman analysis suggests a good agreement between manual and AI analysis of vertical palpebral aperture (94.4% of measurements falling within 2 SDs of the mean). Correlation between the 2 methods yielded a Pearson’s r(126) = 0.87 (P < 0.01) and r2 = 0.76. This feasibility study suggests that existing, open-source approaches to facial analysis can be applied to the clinical assessment of patients with abnormal lid position. The approach could be extended to further quantify clinical assessment of oculoplastic conditions.

Real-time skin chromophore estimation from hyperspectral images using a neural network.

Hyperspectral imaging for in vivo human skin study has shown great potential by providing non-invasive measurement from which information usually invisible to the human eye can be revealed. In particular, maps of skin parameters including oxygen rate, blood volume fraction, and melanin concentration can be estimated from a hyperspectral image by using an optical model and an optimization algorithm. These applications, relying on hyperspectral images acquired with a high-resolution camera especially dedicated to skin measurement, have yielded promising results. However, the data analysis process is relatively expensive in terms of computation cost, with calculation of full-face skin property maps requiring up to 5hours for 3-megapixels hyperspectral images. Such a computation time prevents punctual previewing and quality assessment of the maps immediately after acquisition. To address this issue, we have implemented a neural network that models the optimization-based analysis algorithm. This neural network has been trained on a set of hyperspectral images, acquired from 204 patients and their corresponding skin parameter maps, which were calculated by optimization. The neural network is able to generate skin parameter maps that are visually very faithful to the reference maps much more quickly than the optimization-based algorithm, with computation times as short as 2seconds for a 3-megapixel image representing a full face and 0.5seconds for a 1-megapixel image representing a smaller area of skin. The average deviation calculated on selected areas shows the network's promising generalization ability, even on wide-field full-face images. Currently, the network is adequate for preview purposes, providing relatively accurate results in a few seconds.

Benchmarking commercial emotion detection systems using realistic distortions of facial image datasets

Currently, there are several widely used commercial cloud-based services that attempt to recognize an individual’s emotions based on their facial expressions. Most research into facial emotion recognition has used high-resolution, front-oriented, full-face images. However, when images are collected in naturalistic settings (e.g., using smartphone’s frontal camera), these images are likely to be far from ideal due to camera positioning, lighting conditions, and camera shake. The impact these conditions have on the accuracy of commercial emotion recognition services has not been studied in full detail. To fill this gap, we selected five prominent commercial emotion recognition systems—Amazon Rekognition, Baidu Research, Face++, Microsoft Azure, and Affectiva—and evaluated their performance via two experiments. In Experiment 1, we compared the systems’ accuracy at classifying images drawn from three standardized facial expression databases. In Experiment 2, we first identified several common scenarios (e.g., partially visible face) that can lead to poor-quality pictures during smartphone use, and manipulated the same set of images used in Experiment 1 to simulate these scenarios. We used the manipulated images to again compare the systems’ classification performance, finding that the systems varied in how well they handled manipulated images that simulate realistic image distortion. Based on our findings, we offer recommendations for developers and researchers who would like to use commercial facial emotion recognition technologies in their applications.

The Potential for Undue Patient Exposure during the Use of Telementoring Technology

BackgroundSurgical telementoring holds great promise for safe and effective patient care and medical education, but recording and streaming audio and video introduces the potential for exposure of patient information. Physicians maintain an ethical responsibility to protect the privacy of patients, and privacy violations may carry significant legal liability. Despite the legal treatment of violations as discrete, methods for quantifying and characterizing the exposure of patient information during procedural recordings are lacking. This study is the first to quantify the potential risk for violation of privacy when using a wearable, telementoring technology capable of video and audio recording during surgical procedures in various locations including the operating room, interventional radiology suite, and the intensive care unit.MethodsA head-mounted recording device, Google Glass™, was used to record routine neurosurgical and critical care procedures in a convenience sample of patients. Periods of maximal risk, including the beginning of procedures, were targeted. Recordings were manually coded for discrete instances of exposure of directly identifying information and indirectly identifying information.ResultsTwenty-two procedures were recorded for a total of 12 hours, during which 807 directly identifiable exposures were found. The overall average rate of exposure was 1.13 exposures per minute. Most exposures were full-face images (90%), names (7%), or phone numbers (3%). Indirectly identifying exposures were found to be tattoos, genitals, and caretaker names. The rate of exposures was found to be lower in the operating room (OR) when compared to the intensive care unit (ICU) or interventional radiology (IR) suite (p = 0.0376).ConclusionsHigh rates of potential privacy violations were discovered and found to be related the location of the procedure. Sterile draping of the face prior to recording, when appropriate, would mitigate most exposure risk, though patient names and unique tattoos may be an underappreciated source of potential exposure. This study establishes the most conservative baseline to compare techniques for preventing exposure of patient information on telementoring or video/audio recording/streaming platforms.

Detecting Affect States Using VGG16, ResNet50 and SE-ResNet50 Networks

Affect detection is a key component in developing intelligent human computer interface systems. State-of-the-art affect detection systems assume the availability of full un-occluded face images. This work uses convolutional neural networks with transfer learning to detect 7 basic affect states, viz. Angry, Contempt, Disgust, Fear, Happy and Sad. The paper compares three pre-trained networks, viz. VGG16, ResNet50 and a SE-ResNet50, in which a new architectural block of squeeze and excitation has been integrated with ResNet50. Modified VGG-16, ResNet50 and SE-ResNet50 networks are trained on images from the dataset, and the results are compared. We have been able to achieve validation accuracies of 96.8%, 99.47%, and 97.34% for VGG16, ResNet50 and SE-ResNet50, respectively. Apart from accuracy, the other performance matrices used in this work are precision and recall. Our evaluation, based on these performance matrices, shows that accurate affect detection is obtained from all the three networks with Resnet50 being the most accurate.

A hybrid method for accurate iris segmentation on at-a-distance visible-wavelength images

This work describes a new hybrid method for accurate iris segmentation from full-face images independently of the ethnicity of the subject. It is based on a combination of three methods: facial key-point detection, integro-differential operator (IDO) and mathematical morphology. First, facial landmarks are extracted by means of the Chehra algorithm in order to obtain the eye location. Then, the IDO is applied to the extracted sub-image containing only the eye in order to locate the iris. Once the iris is located, a series of mathematical morphological operations is performed in order to accurately segment it. Results are obtained and compared among four different ethnicities (Asian, Black, Latino and White) as well as with two other iris segmentation algorithms. In addition, robustness against rotation, blurring and noise is also assessed. Our method obtains state-of-the-art performance and shows itself robust with small amounts of blur, noise and/or rotation. Furthermore, it is fast, accurate, and its code is publicly available.

New image analysis tool for facial pore characterization and assessment.

Visible facial pores are an important cosmetic concern especially among young females. Number of different methodologies is used today to assess facial pores and efficacy of technologies. Main limitations of these methods are, however, (a) moderate correlation with visual perception, (b) characterization is mostly limited to size or number of pores, (c) measurement is limited to a smaller area of face, and (d) operational difficulties. In order to address these limitations, we developed a 2D image analysis tool to assess and characterize visible facial pores. Two clinical studies were conducted with northeast Asian skin type females. In the first study, 40 subjects age between 20 and 40 with different degree of pore severity were recruited. In the second study, 15 subjects age between 20 and 40 with enlarged pores were recruited to evaluate pore product efficacy. In both studies, full face images were taken using Visia-CR and assessed by means of the newly developed tool and visual grading. A high correlation between visual grading and pore size was obtained (r=0.86). New methodology was able to differentiate products similar to visual grading. Novel pore image analysis method using 2D skin surface imaging with standard photography has been developed and validated. In addition to pore size measurements, we propose this method to be used to measure pore shape, color, and orientation for a comprehensive characterization of facial pores.

MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation.

Learning-based methods are believed to work well for unconstrained gaze estimation, i.e. gaze estimation from a monocular RGB camera without assumptions regarding user, environment, or camera. However, current gaze datasets were collected under laboratory conditions and methods were not evaluated across multiple datasets. Our work makes three contributions towards addressing these limitations. First, we present the MPIIGaze dataset, which contains 213,659 full face images and corresponding ground-truth gaze positions collected from 15 users during everyday laptop use over several months. An experience sampling approach ensured continuous gaze and head poses and realistic variation in eye appearance and illumination. To facilitate cross-dataset evaluations, 37,667 images were manually annotated with eye corners, mouth corners, and pupil centres. Second, we present an extensive evaluation of state-of-the-art gaze estimation methods on three current datasets, including MPIIGaze. We study key challenges including target gaze range, illumination conditions, and facial appearance variation. We show that image resolution and the use of both eyes affect gaze estimation performance, while head pose and pupil centre information are less informative. Finally, we propose GazeNet, the first deep appearance-based gaze estimation method. GazeNet improves on the state of the art by 22 percent (from a mean error of 13.9 degrees to 10.8 degrees) for the most challenging cross-dataset evaluation.

Perception of Smile Esthetics by Orthodontists and Laypersons: Full Face and A Localized View of The Social and Spontaneous Smiles.

The aim of the study was to evaluate the parameters that might affect the esthetic perception of localized and full-face views of social and spontaneous smiles. Video records of 40 individuals were used. Further, 200 images of each individual were captured for social and spontaneous smiles with and without calibration glasses. Full-face images of social smile (SSF) and spontaneous smile (smile of joy) (JSF) were obtained. Furthermore, the mouth-area images of the same social (SSM) and spontaneous (JSM) smiles from each subject were acquired. Here 160 images were evaluated by orthodontists and laypersons using the Q-sort method. The data were analyzed with logistic regression and independent samples t-test. From the orthodontic perspective, upper lip thickness (p=0.004), lip curtain over incisors (p=0.016), maxillary incisor display (p=0.01), and buccal corridor ratio (p=0.006) were significant to determine attractive and unattractive images when viewing localized social smiles. Laypersons identified no particular parameter to explain the variation in preferences for all the image groups (p>0.05). Upper lip thickness and maxillary incisor display during smiling were found to be effective for distinguishing images as attractive and unattractive by orthodontists. No objective evaluations for the attractiveness were made by laypersons.