The Auto-eFACE: Machine Learning-Enhanced Program Yields Automated Facial Palsy Assessment Tool

Correlation of 3D Morphometric Changes, Kinematics, and Muscle Activity During Smile

OBJECTIVE

Knowing the morphological, kinematic, and electrophysiological parameters of the smile in healthy individuals may contribute to evaluating, planning, and monitoring the smile reanimation. This study aimed to determine the correlation between 3D morphometric changes, movement kinematics, and muscle activity in the facial soft tissue of healthy individuals.

METHOD

In this cohort study, 20 volunteers were selected from healthy individuals with no facial disorders. During smiling, three-dimensional face scanning, facial motion capture, and surface electromyography (sEMG) were performed. The average displacement, velocity, and acceleration during facial movements were measured. The mean change in 3D surface morphometry and activation of the zygomaticus major were determined.

RESULTS

The volunteers, comprising 10 males and 10 females, had a mean age of 24 ± 10 years; for female, mean age was 23 ± 5 years and for men 26 ± 13 years. Significant correlations were found between kinematic and morphometric data (r = 0.51, p < 0.001), sEMG and morphometric (r = 0.50, p < 0.001) data, and sEMG and kinematic data (r = 0.49, p < 0.002). The maximum acceleration occurred during approximately 65% of the muscle activation time and 64% of the peak muscle activation value. Additionally, the maximum velocity was reached at around 73% of the muscle activation time and 67% of the peak muscle activation value. Furthermore, the maximum displacement values were observed at approximately 88% of the muscle activation time and 76% of the peak muscle activation value.

CONCLUSION

The findings may provide insights into the smile's functional parameters, contribute to understanding facial muscle-related disorders, and aid in improving the diagnosis and treatment of the smile.

LEVEL OF EVIDENCE

NA Laryngoscope, 134:3112-3119, 2024.

Analysis of factors associated with patients' global impression of change scores in inpatients with peripheral facial palsy: A retrospective study at a Korean Medicine Hospital

The incidence of peripheral facial palsy is on the rise, with psychological issues influencing quality of life due to visible symptoms. Consequently, patient-reported outcome measures are critical in assessing the disease burden and the treatment efficacy of peripheral facial palsy from patients' perspective. This study examines factors influencing patients' global impression of change (PGIC) scores, a type of patient-reported outcome measure, in hospitalized patients with peripheral facial palsy. This retrospective study analyzed the electronic medical records of 200 patients with peripheral facial palsy who were admitted to the Korean Medicine Hospital, Kyung Hee University Medical Center from May 1, 2022 to April 30, 2023. Impact of demographic, electrophysiological, intervention, and clinical factors on PGIC scores were evaluated. Descriptive statistics showed that the length of hospitalization (P = .020), time from disease onset to hospitalization (P = .022), lacrimal disorders (P = .002), House-Brackmann (HB) grade evaluated at admission (P = .016) and at discharge (P < .001), improvement in HB-grade from admission to discharge (P = .002), and total facial disability index (FDI) score at discharge (P < .001) were significantly associated with PGIC scores. In multivariate logistic regression analysis, HB-grade at admission (OR: 13.89, 95% CI: 2.18-113.60), length of stay (OR: 0.27, 95% CI: 0.07-0.92), time from disease onset to hospitalization (OR: 5.55, 95% CI: 1.36-24. 77), tear-related symptoms (OR: 0.41, 95% CI: 0.17-0.96), total FDI score (OR: 0.45, 95% CI: 0.20-0.98), and greater improvement in HB-grade at discharge compared to admission (OR: 0.08, 95% CI: 0.02-0.31) were significantly associated with PGIC scores. Patients with milder initial disease severity, hospitalization period exceeding 7 days, shorter time from disease onset to hospitalization, improvement of lacrimal symptoms, total FDI score, and HB-grade between admission and discharge experienced more significant subjective improvement in peripheral facial palsy.

Is pre-operative electromyography a reliable tool in differentiating acute and chronic facial palsy? A preliminary evaluation in patients treated with triple innervation facial reanimation

Electromyographic evaluation is a reliable tool for confirming facial palsy and assessing its severity. It allows differentiating facial paresis and paralysis, and further distinguishes acute palsies, still showing muscle fibrillations, from chronic cases. This article aims to show that EMG fibrillations might represent a better criterion to differentiate acute and chronic palsies than the standard 18-24 months' cut-off usually employed for classification and treatment purposes. We performed a cohort study using the eFACE tool for comparing triple innervation facial reanimation results in patients with EMG fibrillation treated <12 months, 12-18 months, and >18 months from paralysis onset. Patients showed a statistically significant post-operative improvement in all eFACE items, both in the whole sample and in the three groups. Only the deviation from the optimal score for the gentle eye closure item in group 2 didn't reach statistical significance (p = 0.173). The post-operative results were comparable in the three groups, as the Kruskal-Wallis test showed a difference only for the platysmal synkinesis item scores, which were significantly lower in group 3 (p = 0.025).

Comparison of Objective Facial Metrics on Both Sides of the Face Among Patients with Severe Bell's Palsy Treated with Mirror Effect Plus Protocol Rehabilitation Versus Controls

Objective: The extent to which the healthy hemiface dynamically contributes to facial synchronization during facial rehabilitation has been largely unstudied. This study compares the synchronization of both hemifaces in severe Bell's palsy patients who either received facial rehabilitation called "Mirror Effect Plus Protocol" (MEPP) or basic counseling. Methods: Baseline and 1-year postonset data from 39 patients (19 = MEPP and 20 = basic counseling) were retrospectively analyzed using Emotrics+, a software that generates facial metrics with artificial intelligence (AI) algorithms. Paired t-tests were used for intrasubject comparisons of hemifaces, and mixed model analysis were used to compare between groups. Results: For voluntary movements, a significant difference in favor of the MEPP group was only found for smiling (p = 0.025*). However, at 1-year postonset, the control group showed significant variability between hemifaces for most synkinesis measurements [nasolabial fold (p = 0.029*); eye area (p = 0.043*); palpebral fissure (p = 0.011*)]. Conclusion: In this study, a better synchronization of both hemifaces was found in the MEPP group. Interestingly, motor adaptation in movement amplitude of the healthy hemiface seemed to contribute to this synchronization in MEPP patients. Further studies are needed to standardize the procedure of AI measurements and to adapt it for clinical use.

Toward an application of automatic evaluation system for central facial palsy using two simple evaluation indices in emergency medicine

A stroke is a medical emergency and thus requires immediate treatment. Paramedics should accurately assess suspected stroke patients and promptly transport them to a hospital with stroke care facilities; however, current assessment procedures rely on subjective visual assessment. We aim to develop an automatic evaluation system for central facial palsy (CFP) that uses RGB cameras installed in an ambulance. This paper presents two evaluation indices, namely the symmetry of mouth movement and the difference in mouth shape, respectively, extracted from video frames. These evaluation indices allow us to quantitatively evaluate the degree of facial palsy. A classification model based on these indices can discriminate patients with CFP. The results of experiments using our dataset show that the values of the two evaluation indices are significantly different between healthy subjects and CFP patients. Furthermore, our classification model achieved an area under the curve of 0.847. This study demonstrates that the proposed automatic evaluation system has great potential for quantitatively assessing CFP patients based on two evaluation indices.

Validation of the Spanish version of the Electronic Facial Palsy Assessment (eFACE)

PURPOSE

The clinician-graded electronic facial paralysis assessment (eFACE) is a relatively new digital tool for assessing facial palsy. The present study aimed to determine the validity and reliability of the Spanish version of the eFACE.

METHODS

Forward-backward translation from the original English version was performed. Videos and photographs from 65 adult patients with unilateral facial paralysis (any severity, time course, and etiology) were evaluated twice by five otolaryngologists with varying levels of experience in facial palsy evaluation. Internal consistency was measured using Cronbach's α and the intra- and inter-rater reliability were measured using intraclass correlation coefficient. Concurrent validity was established by calculating Spearman's rho correlation (ρ) between the eFACE and the House-Brackmann scale (H-B) and Pearson's correlation (r) between the eFACE and the Sunnybrook Facial Grading System (SFGS).

RESULTS

The Spanish version of the eFACE showed good internal consistency (Cronbach's α > 0.8). The intra-rater reliability was nearly perfect for the total score (intraclass correlation coefficient: 0.95-0.99), static score (0.92-0.96), and dynamic score (0.96-0.99) and important-to-excellent for synkinesis score (0.79-0.96). The inter-rater reliability was excellent for the total score (0.85-0.93), static score (0.80-0.90), and dynamic score (0.90-0.95) and moderate-to-important for the synkinesis score (0.55-0.78). The eFACE had a very strong correlation with the H-B (ρ =  - 0.88 and  - 0.85 for each evaluation, p < 0.001) and the SFGS (r = 0.92 and 0.91 each evaluation, p < 0.001).

CONCLUSION

The Spanish version of the eFACE is a reliable and valid instrument for assessment of facial function in the diagnosis and treatment of patients with facial paralysis.

Evaluation of the Patient with Facial Palsy: A Multidisciplinary Approach

This article aims to provide an overview of the management of facial palsy within a multidisciplinary team setting and discusses considerations used to develop patient-specific management plans. The national landscape of facial function services is also discussed including suggestions on what may enable a more equitable and sustainable service for the future.

"Facekit"-Toward an Automated Facial Analysis App Using a Machine Learning-Derived Facial Recognition Algorithm

Introduction: Multiple tools have been developed for facial feature measurements and analysis using facial recognition machine learning techniques. However, several challenges remain before these will be useful in the clinical context for reconstructive and aesthetic plastic surgery. Smartphone-based applications utilizing open-access machine learning tools can be rapidly developed, deployed, and tested for use in clinical settings. This research compares a smartphone-based facial recognition algorithm to direct and digital measurement performance for use in facial analysis. Methods: Facekit is a camera application developed for Android that utilizes ML Kit, an open-access computer vision Application Programing Interface developed by Google. Using the facial landmark module, we measured 4 facial proportions in 15 healthy subjects and compared them to direct surface and digital measurements using intraclass correlation (ICC) and Pearson correlation. Results: Measurement of the naso-facial proportion achieved the highest ICC of 0.321, where ICC > 0.75 is considered an excellent agreement between methods. Repeated measures analysis of variance of proportion measurements between ML Kit, direct and digital methods, were significantly different (F[2,14] = 6-26, P<<.05). Facekit measurements of orbital, orbitonasal, naso-oral, and naso-facial ratios had overall low correlation and agreement to both direct and digital measurements (R<<0.5, ICC<<0.75). Conclusion: Facekit is a smartphone camera application for rapid facial feature analysis. Agreement between Facekit's machine learning measurements and direct and digital measurements was low. We conclude that the chosen pretrained facial recognition software is not accurate enough for conducting a clinically useful facial analysis. Custom models trained on accurate and clinically relevant landmarks may provide better performance.

The Emerging Role of Artificial Intelligence Tools for Outcome Measurement in Facial Reanimation Surgery: A Review

Importance: Surgeons treating facial paralysis with reanimation surgery measure the outcomes of surgery and adjust treatment to each patient's needs. Our objective is to review the current subjective facial paralysis assessment tools and the emerging computer-based objective analysis, which may involve artificial intelligence. Observations: In recent years, many new automated approaches to outcome measurement in facial reanimation surgery have been developed. Most of these tools utilize artificial intelligence to analyze emotional expression and symmetry of facial landmarks. Other tools have provided automated approaches to existing clinician-guided scales. Conclusions: Newly developed computer-based tools using artificial intelligence have been developed to both improve existing clinician-graded scales and provide new approaches to facial symmetry and emotional expressivity analysis.

Commentary on: "Chemodenervation Algorithm: Functional and Aesthetic Considerations for Facial Harmony in Patients with Post-Facial Paralysis Synkinesis," by Hetzler et al

In this commentary, we discuss Hetzler et al.'s article, "Chemodenervation Algorithm: Functional and Aesthetic Considerations for Facial Harmony in Patients with Post-Facial Paralysis Synkinesis." The authors do an excellent job of presenting a guide for practitioners to use when initiating chemodenervation treatment for patients with nonflaccid facial paralysis. Standardization of outcome assessment tools and rigorous data collection will further refine treatment algorithms.

Deep Learning for the Assessment of Facial Nerve Palsy: Opportunities and Challenges

Automated evaluation of facial palsy using machine learning offers a promising solution to the limitations of current assessment methods, which can be time-consuming, labor-intensive, and subject to clinician bias. Deep learning-driven systems have the potential to rapidly triage patients with varying levels of palsy severity and accurately track recovery over time. However, developing a clinically usable tool faces several challenges, such as data quality, inherent biases in machine learning algorithms, and explainability of decision-making processes. The development of the eFACE scale and its associated software has improved clinician scoring of facial palsy. Additionally, Emotrics is a semiautomated tool that provides quantitative data of facial landmarks on patient photographs. The ideal artificial intelligence (AI)-enabled system would analyze patient videos in real time, extracting anatomic landmark data to quantify symmetry and movement, and estimate clinical eFACE scores. This would not replace clinician eFACE scoring but would offer a rapid automated estimate of both anatomic data, similar to Emotrics, and clinical severity, similar to the eFACE. This review explores the current state of facial palsy assessment, recent advancements in AI, and the opportunities and challenges in developing an AI-driven solution.

Identification of Stable Anatomical Landmarks During Facial Expressions

Background: Identifying stable anatomical landmarks during facial expressions is necessary to register and align three-dimensional (3D) data, determine the common origin and motion axis, and calculate displacement, velocity, and acceleration of relative motion. Objective: Our study aimed to determine the stable anatomical landmarks during facial expressions. Methods: We included 30 volunteers in our study and asked them to perform resting, mouth opening, showing teeth, clenching teeth, eye closure, smiling, whistling, eyebrow raising, and disgusted facial expressions. We recorded the 3D movements of passive reflective markers with optoelectronic cameras, which we attached to 10 bilateral and 8 midline landmarks. Results: We determined that the inner corner of the eye while mouth opening, the hairline superior while showing teeth, the infraorbital rim while clenching teeth, the alare while eye closure, the inner corner of the eye while smiling, the mideyebrow while whistling, corners of the mouth while eyebrow raising, and hairline superior while disgust are the most stable anatomical landmarks. Conclusions: Our study identified immobile soft tissue landmarks specific to facial expression.

Evidence-Based Practices in Facial Reanimation Surgery

LEARNING OBJECTIVES

After studying this article, the participant should be able to: 1. Describe the causes and preoperative evaluation of facial paralysis. 2. Discuss techniques to restore corneal sensation and eyelid closure, elevation of the upper lip for smile, and depression of the lower lip for lip symmetry. 3. Outline treatment goals, surgical treatment options, timing of repair, and other patient-specific considerations in appropriate technique selection.

SUMMARY

Congenital facial paralysis affects 2.7 per 100,000 children; Bell palsy affects 23 per 100,000 people annually; and even more people are affected when considering all other causes. Conditions that impair facial mimetics impact patients' social functioning and emotional well-being. Dynamic and static reconstructive methods may be used individually or in concert to achieve adequate blink restoration, smile strength and spontaneity, and lower lip depression. Timing of injury and repair, patient characteristics such as age, and cause of facial paralysis are all considered in selecting the most appropriate reconstructive approach. This article describes evidence-based management of facial paralysis.

Embellishing Emotrics for a More Complete Emotion Analysis: Addition of the Nasolabial Fold

Background: The nasolabial fold (NLF) greatly contributes to facial aesthetics; changes to NLF depth and vector are disfiguring in patients with facial paralysis (FP). NLF parameters are integral to clinician-graded outcomes, but automated programs currently lack NLF identification capabilities. Objective: To incorporate an automated NLF identification and quantification function into the facial landmark program, Emotrics, and to compare new Emotrics-derived NLF data to clinician-graded electronic facial paralysis assessment (eFACE) data for accuracy. Methods: Photographs of 135 patients with FP were marked bilaterally, using identification markers manually placed on each NLF. A machine learning model was trained to automatically localize the markers using these data. Once Emotrics accurately identified the NLF and its corresponding vector, photographs of 20 additional patients who underwent facial reanimation procedures were assessed by the algorithm. Results: The enhanced Emotrics algorithm successfully identified the NLF, and measured the vector from midline, in a series of patients with FP. NLF vector data closely matched corresponding eFACE parameters. Furthermore, changes in NLF presence and vector were detected following facial reanimation procedures. Conclusion: The Emotrics program now provides critical NLF data, providing objective parameters for clinicians interested in changing NLF dynamics after FP.

Evidence-Based Medicine: Facial Reanimation

This article provides a brief historical overview of the assessment instruments that have been developed to categorize the severity of disease in patients with facial palsy. Important advances in the quality of these instruments are discussed. The modern-day instruments that are commonly required for evidence-based patient assessment are then presented, with emphasis on the level of evidence of the studies that describe these instruments.

The evolution of facial reanimation techniques

This review article provides an updated discussion on evidence-based practices related to the evaluation and management of facial paralysis. Ultimately, the goals of facial reanimation include obtaining facial symmetry at rest, providing corneal protection, restoring smile symmetry and facial movement for functional and aesthetic purposes. The treatment of facial nerve injury is highly individualized, especially given the wide heterogeneity regarding the degree of initial neuronal insult and eventual functional outcome. Recent advancements in facial reanimation techniques have better equipped clinicians to approach challenging patient scenarios with reliable, effective strategies. We discuss how technology such as machine learning software has revolutionized pre- and post-intervention assessments and provide an overview of current controversies including timing of intervention, choice of donor nerve, and management of nonflaccid facial palsy with synkinesis. We highlight novel considerations to mainstay conservative management strategies and examine innovations in modern surgical techniques with a focus on gracilis free muscle transfer. Innervation sources, procedural staging, coaptation patterns, and multi-vector and multi-muscle paddle design are modifications that have significantly evolved over the past decade.

DeepSmile: Anomaly Detection Software for Facial Movement Assessment

Facial movements are crucial for human interaction because they provide relevant information on verbal and non-verbal communication and social interactions. From a clinical point of view, the analysis of facial movements is important for diagnosis, follow-up, drug therapy, and surgical treatment. Current methods of assessing facial palsy are either (i) objective but inaccurate, (ii) subjective and, thus, depending on the clinician's level of experience, or (iii) based on static data. To address the aforementioned problems, we implemented a deep learning algorithm to assess facial movements during smiling. Such a model was trained on a dataset that contains healthy smiles only following an anomaly detection strategy. Generally speaking, the degree of anomaly is computed by comparing the model's suggested healthy smile with the person's actual smile. The experimentation showed that the model successfully computed a high degree of anomaly when assessing the patients' smiles. Furthermore, a graphical user interface was developed to test its practical usage in a clinical routine. In conclusion, we present a deep learning model, implemented on open-source software, designed to help clinicians to assess facial movements.

Comprehensive assessment of facial paralysis based on facial animation units

Quantitative grading and classification of the severity of facial paralysis (FP) are important for selecting the treatment plan and detecting subtle improvement that cannot be detected clinically. To date, none of the available FP grading systems have gained widespread clinical acceptance. The work presented here describes the development and testing of a system for FP grading and assessment which is part of a comprehensive evaluation system for FP. The system is based on the Kinect v2 hardware and the accompanying software SDK 2.0 in extracting the real time facial landmarks and facial animation units (FAUs). The aim of this paper is to describe the development and testing of the FP assessment phase (first phase) of a larger comprehensive evaluation system of FP. The system includes two phases; FP assessment and FP classification. A dataset of 375 records from 13 unilateral FP patients was compiled for this study. The FP assessment includes three separate modules. One module is the symmetry assessment of both facial sides at rest and while performing five voluntary facial movements. Another module is responsible for recognizing the facial movements. The last module assesses the performance of each facial movement for both sides of the face depending on the involved FAUs. The study validates that the FAUs captured using the Kinect sensor can be processed and used to develop an effective tool for the automatic evaluation of FP. The developed FP grading system provides a detailed quantitative report and has significant advantages over the existing grading scales. It is fast, easy to use, user-independent, low cost, quantitative, and automated and hence it is suitable to be used as a clinical tool.

Pathogenesis, diagnosis and therapy of facial synkinesis: A systematic review and clinical practice recommendations by the international head and neck scientific group

Introduction

Post-paralytic facial synkinesis after facial nerve injury produces functional disabilities and mimetic deficits, but also cosmetic and non-motor psychosocial impairments for the patients. These patients typically have a high and continuous high motivation for rehabilitation. The aim is to inform the affected patients and their therapeutic professionals (otorhinolaryngologist - head and neck surgeons; oral-maxillofacial surgeons, plastic and reconstructive surgeons, neurosurgeons, neurologists, and mime therapists be it speech and language therapy- or physiotherapy-based) and to provide practical recommendations for diagnostics and a stepwise systematic treatment approach of facial synkinesis.

Methods

In the first phase, a systematic literature search on the topic in PubMed and ScienceDirect starting in 2008 resulted in 132 articles. These were the basis for the review and a comprehensive series of consensus statements on the most important diagnostic tests and treatment options. In the second phase, one consensus article circulated among the membership of the International Head and Neck Scientific Group until a final agreement was reached for all recommendations.

Results

Diagnostics should include a standardized assessment of the degree of synkinesis using validated clinician-graded instruments and synkinesis-specific patient-reported outcome measures. Treatments for facial synkinesis include facial training mainly based on facial biofeedback retraining, chemodenervation with botulinum toxin, selective neurectomy, myectomy, and any combination treatment of these options.

Conclusion

A basic understanding of the pathomechanisms of synkinesis is essential to understand the treatment strategies. A standardized assessment of the synkinetic symptoms and the individual synkinesis pattern is needed. The first-line treatment is facial training, followed by botulinum toxin. Surgery is reserved for individual cases with unsatisfactory first-line treatment.

A Ready-to-Use Grading Tool for Facial Palsy Examiners—Automated Grading System in Facial Palsy Patients Made Easy

Background: The grading process in facial palsy (FP) patients is crucial for time- and cost-effective therapy decision-making. The House-Brackmann scale (HBS) represents the most commonly used classification system in FP diagnostics. This study investigated the benefits of linking machine learning (ML) techniques with the HBS. Methods: Image datasets of 51 patients seen at the Department of Plastic, Hand, and Reconstructive Surgery at the University Hospital Regensburg, Germany, between June 2020 and May 2021, were used to build the neural network. A total of nine facial poses per patient were used to automatically determine the HBS. Results: The algorithm had an accuracy of 98%. The algorithm processed the real patient image series (i.e., nine images per patient) in 112 ms. For optimized accuracy, we found 30 training runs to be the most effective training length. Conclusion: We have developed an easy-to-use, time- and cost-efficient algorithm that provides highly accurate automated grading of FP patient images. In combination with our application, the algorithm may facilitate the FP surgeon’s clinical workflow.

Towards a Reliable and Rapid Automated Grading System in Facial Palsy Patients: Facial Palsy Surgery Meets Computer Science

Background: Reliable, time- and cost-effective, and clinician-friendly diagnostic tools are cornerstones in facial palsy (FP) patient management. Different automated FP grading systems have been developed but revealed persisting downsides such as insufficient accuracy and cost-intensive hardware. We aimed to overcome these barriers and programmed an automated grading system for FP patients utilizing the House and Brackmann scale (HBS). Methods: Image datasets of 86 patients seen at the Department of Plastic, Hand, and Reconstructive Surgery at the University Hospital Regensburg, Germany, between June 2017 and May 2021, were used to train the neural network and evaluate its accuracy. Nine facial poses per patient were analyzed by the algorithm. Results: The algorithm showed an accuracy of 100%. Oversampling did not result in altered outcomes, while the direct form displayed superior accuracy levels when compared to the modular classification form (n = 86; 100% vs. 99%). The Early Fusion technique was linked to improved accuracy outcomes in comparison to the Late Fusion and sequential method (n = 86; 100% vs. 96% vs. 97%). Conclusions: Our automated FP grading system combines high-level accuracy with cost- and time-effectiveness. Our algorithm may accelerate the grading process in FP patients and facilitate the FP surgeon’s workflow.

Is There a Difference in Facial Emotion Recognition after Stroke with vs. without Central Facial Paresis?

The Facial Feedback Hypothesis (FFH) states that facial emotion recognition is based on the imitation of facial emotional expressions and the processing of physiological feedback. In the light of limited and contradictory evidence, this hypothesis is still being debated. Therefore, in the present study, emotion recognition was tested in patients with central facial paresis after stroke. Performance in facial vs. auditory emotion recognition was assessed in patients with vs. without facial paresis. The accuracy of objective facial emotion recognition was significantly lower in patients with vs. without facial paresis and also in comparison to healthy controls. Moreover, for patients with facial paresis, the accuracy measure for facial emotion recognition was significantly worse than that for auditory emotion recognition. Finally, in patients with facial paresis, the subjective judgements of their own facial emotion recognition abilities differed strongly from their objective performances. This pattern of results demonstrates a specific deficit in facial emotion recognition in central facial paresis and thus provides support for the FFH and points out certain effects of stroke.

Review on Facial-Recognition-Based Applications in Disease Diagnosis

Diseases not only manifest as internal structural and functional abnormalities, but also have facial characteristics and appearance deformities. Specific facial phenotypes are potential diagnostic markers, especially for endocrine and metabolic syndromes, genetic disorders, facial neuromuscular diseases, etc. The technology of facial recognition (FR) has been developed for more than a half century, but research in automated identification applied in clinical medicine has exploded only in the last decade. Artificial-intelligence-based FR has been found to have superior performance in diagnosis of diseases. This interdisciplinary field is promising for the optimization of the screening and diagnosis process and assisting in clinical evaluation and decision-making. However, only a few instances have been translated to practical use, and there is need of an overview for integration and future perspectives. This review mainly focuses on the leading edge of technology and applications in varieties of disease, and discusses implications for further exploration.

Facial Emotion Recognition in Patients with Post-Paralytic Facial Synkinesis—A Present Competence

Facial palsy is a movement disorder with impacts on verbal and nonverbal communication. The aim of this study is to investigate the effects of post-paralytic facial synkinesis on facial emotion recognition. In a prospective cross-sectional study, we compared facial emotion recognition between n = 30 patients with post-paralytic facial synkinesis (mean disease time: 1581 ± 1237 days) and n = 30 healthy controls matched in sex, age, and education level. Facial emotion recognition was measured by the Myfacetraining Program. As an intra-individual control condition, auditory emotion recognition was assessed via Montreal Affective Voices. Moreover, self-assessed emotion recognition was studied with questionnaires. In facial as well as auditory emotion recognition, on average, there was no significant difference between patients and healthy controls. The outcomes of the measurements as well as the self-reports were comparable between patients and healthy controls. In contrast to previous studies in patients with peripheral and central facial palsy, these results indicate unimpaired ability for facial emotion recognition. Only in single patients with pronounced facial asymmetry and severe facial synkinesis was an impaired facial and auditory emotion recognition detected. Further studies should compare emotion recognition in patients with pronounced facial asymmetry in acute and chronic peripheral paralysis and central and peripheral facial palsy.

Towards Facial Gesture Recognition in Photographs of Patients with Facial Palsy

Humans express their emotions verbally and through actions, and hence emotions play a fundamental role in facial expressions and body gestures. Facial expression recognition is a popular topic in security, healthcare, entertainment, advertisement, education, and robotics. Detecting facial expressions via gesture recognition is a complex and challenging problem, especially in persons who suffer face impairments, such as patients with facial paralysis. Facial palsy or paralysis refers to the incapacity to move the facial muscles on one or both sides of the face. This work proposes a methodology based on neural networks and handcrafted features to recognize six gestures in patients with facial palsy. The proposed facial palsy gesture recognition system is designed and evaluated on a publicly available database with good results as a first attempt to perform this task in the medical field. We conclude that, to recognize facial gestures in patients with facial paralysis, the severity of the damage has to be considered because paralyzed organs exhibit different behavior than do healthy ones, and any recognition system must be capable of discerning these behaviors.

Multidisciplinary Care of Patients with Facial Palsy: Treatment of 1220 Patients in a German Facial Nerve Center

To determine treatment and outcome in a tertiary multidisciplinary facial nerve center, a retrospective observational study was performed of all patients referred between 2007 and 2018. Facial grading with the Stennert index, the Facial Clinimetric Evaluation (FaCE) scale, and the Facial Disability Index (FDI) were used for outcome evaluation; 1220 patients (58.4% female, median age: 50 years; chronic palsy: 42.8%) were included. Patients with acute and chronic facial palsy were treated in the center for a median of 3.6 months and 10.8 months, respectively. Dominant treatment in the acute phase was glucocorticoids ± acyclovir (47.2%), followed by a significant improvement of all outcome measures (p < 0.001). Facial EMG biofeedback training (21.3%) and botulinum toxin injections (11%) dominated the treatment in the chronic phase, all leading to highly significant improvements according to facial grading, FDI, and FaCE (p < 0.001). Upper eyelid weight (3.8%) and hypoglossal–facial-nerve jump suture (2.5%) were the leading surgical methods, followed by improvement of facial motor function (p < 0.001) and facial-specific quality of life (FDI, FaCE; p < 0.05). A standardized multidisciplinary team approach in a facial nerve center leads to improved facial and emotional function in patients with acute or chronic facial palsy.

Validation of a new graphic facial nerve grading system: FAME scale

Background and Objective:

Most of the existing qualitative facial nerve grading systems are very subjective while the quantitative grading systems are more complex, require longer data input time and specific software. There is a need for having a scoring system with graphic criteria to improve the subjectivity, reliability and convenience. We aimed to develop and validate such a reliable graphic scale for use in Bell's palsy.

Methods:

Face videos of patients with unilateral facial paralysis were recorded using smartphones and analyzed for six items including five voluntary facial movements apart from complications of facial palsy (synkinesis, hyperkinesis, and contracture). 15 videos were used for pilot study, 75 for the development of scale and 110 for its validation. Each video was rated on two separate occasions by 3 independent raters, a score of 0-4 was assigned to each item using the graphic scoring criteria, and a composite score was obtained (range 0–24). Five disease severity categories: normal (score 0), mild (score 1–6), moderate (score 7-12), severe (score: 13–18) and profound facial weakness (score: 19-24).

Results:

The proposed scale and its component items had high inter-rater and intra-rater reliability (Kappa >0.7). Good correlation (Pearson co-efficient >0.7) was seen among the voluntary movements. The proposed scale is a valid tool to score motor deficits and complications of facial palsy.

Conclusions:

The proposed scale is a valid and reliable graphic scale to describe facial motor dysfunction and its secondary defects.

Toward a Universal Measure of Facial Difference Using Two Novel Machine Learning Models

A sensitive, objective, and universally accepted method of measuring facial deformity does not currently exist. Two distinct machine learning methods are described here that produce numerical scores reflecting the level of deformity of a wide variety of facial conditions.

METHODS

The first proposed technique utilizes an object detector based on a cascade function of Haar features. The model was trained using a dataset of 200,000 normal faces, as well as a collection of images devoid of faces. With the model trained to detect normal faces, the face detector confidence score was shown to function as a reliable gauge of facial abnormality. The second technique developed is based on a deep learning architecture of a convolutional autoencoder trained with the same rich dataset of normal faces. Because the convolutional autoencoder regenerates images disposed toward their training dataset (ie, normal faces), we utilized its reconstruction error as an indicator of facial abnormality. Scores generated by both methods were compared with human ratings obtained using a survey of 80 subjects evaluating 60 images depicting a range of facial deformities [rating from 1 (abnormal) to 7 (normal)].

RESULTS

The machine scores were highly correlated to the average human score, with overall Pearson's correlation coefficient exceeding 0.96 (P < 0.00001). Both methods were computationally efficient, reporting results within 3 seconds.

CONCLUSIONS

These models show promise for adaptation into a clinically accessible handheld tool. It is anticipated that ongoing development of this technology will facilitate multicenter collaboration and comparison of outcomes between conditions, techniques, operators, and institutions.

[Diagnostics and Therapy of Idiopathic Facial Palsy (Bell's Palsy)]

The purpose of this review is to report the knowledge for otolaryngologists on standard of care, latest advances, interesting new findings and controversies about the treatment of Bell's palsy. This review is focusing on the acute phase of the disease. The chronic phase, with incomplete, incorrect or no recovery of the palsy, is described briefly. Treatment with prednisolone alone within 72 hours after onset still is the cornerstone of the treatment. The role of antivirals still is unclear. Since 2009 no new and breakthrough clinical trials with influence on the treatment standards have been performed. A study to clarify the role of prednisolone treatment in children is ongoing. Patient-related outcome measures like the Facial Clinimetric Evaluation Scale and the Facial Disability Index are important tools to assess the subjective severity of the disease and psychosocial impact of Bell's palsy next to the motor deficits. Simplified subjective electronic grading systems like the eFACE and first automated image analysis systems have been introduced. Studies clarifying the role of antivirals for severe cases are urgently needed as well as studies on the role of salvage second line therapy after insufficient response to initial corticosteroid treatment. An international consensus on the outcome measures in diagnostics and follow-up is also needed.

Effect of an Intensified Combined Electromyography and Visual Feedback Training on Facial Grading in Patients With Post-paralytic Facial Synkinesis

Background: There is no current standard for facial synkinesis rehabilitation programs. The benefit and stability of effect of an intensified 10-day facial training combining electromyography and visual biofeedback training was evaluated.

Methods: Fifty-four patients (77.8% female; median age: 49.5 years) with post-paralytic facial synkinesis (median time to onset of paralysis: 31.1 months) were included in retrospective longitudinal study between January 2013 and June 2016. Facial function was assesses at baseline (T0), first days of training (T1), last day of training (T2), and follow-up visit (T3) at a median time of 6 months later using the House-Brackmann (HB) facial nerve grading system, Stennert index (SI), Facial Nerve Grading System 2.0 (FNGS 2.0), and Sunnybrook Facial Grading System (SFGS). Pairwise comparisons between the time points with post-hoc Bonferroni correction were performed.

Results: No significant changes of the gradings and subscores were seen between T0 and T1 (all p > 0.01). The 10-day combined and intensified feedback training between T1 and T2 improved facial symmetry and decreased synkinetic activity. Facial grading with the FNGS 2.0 or the SFGS were most suited to depict the training effect. FNGS 2.0, regional score, FNGS 2.0, synkinesis score, and FNGS 2.0 total score improved significantly (all p ≤ 0.0001). Both, the FNGS 2.0 and the SFGS showed the strongest improvement in the nasolabial fold/zygomatic and the oral region. Neither the age of the patient (r = 0.168; p = 0.224), the gender (r = 0.126; p = 0.363) nor the length of the interval between onset of the palsy and training start (r = 0.011; p = 0.886) correlated with the changes of the SFGS between T1 and T2. The results remained stable between T2 and T3 without any further significant change.

Conclusion: Intensified daily combined electromyography and visual biofeedback training over 10 days was effective in patients with facial synkinesis and benefits were stable 6 months after therapy.

Facial paralysis: timing of repair and management of the nonflaccidly paralyzed face

PURPOSE OF REVIEW

To explore recent advances in therapeutic interventions for nonflaccid facial paralysis (NFFP), including new evidence for surgical and nonsurgical treatments. Timing of treatment is also discussed, along with possible future treatments.

RECENT FINDINGS

NFFP remains a difficult disease to treat. Chemodenervation with botulinum toxin remains a first-line treatment to suppress aberrant and antagonistic movements during voluntary use of muscles. More permanent treatments such as selective neurectomy, myectomy, and nerve and muscle transfers have been shown to offer promising results for the nonflaccidly paralyzed face.

SUMMARY

NFFP is commonly seen in patients who have incomplete recovery from facial paralysis, and carries high psychosocial morbidity. A large array of treatments have been described in the literature, both procedural and nonprocedural. Both treatment type and timing are important in optimal patient recovery.

Facial Paralysis Detection on Images Using Key Point Analysis

The inability to move the muscles of the face on one or both sides is known as facial paralysis, which may affect the ability of the patient to speak, blink, swallow saliva, eat, or communicate through natural facial expressions. The well-being of the patient could also be negatively affected. Computer-based systems as a means to detect facial paralysis are important in the development of standardized tools for medical assessment, treatment, and monitoring; additionally, they are expected to provide user-friendly tools for patient monitoring at home. In this work, a methodology to detect facial paralysis in a face photograph is proposed. A system consisting of three modules—facial landmark extraction, facial measure computation, and facial paralysis classification—was designed. Our facial measures aim to identify asymmetry levels within the face elements using facial landmarks, and a binary classifier based on a multi-layer perceptron approach provides an output label. The Weka suite was selected to design the classifier and implement the learning algorithm. Tests on publicly available databases reveal outstanding classification results on images, showing that our methodology that was used to design a binary classifier can be expanded to other databases with great results, even if the participants do not execute similar facial expressions.

Reliability Between In-Person and Still Photograph Assessment of Facial Function in Facial Paralysis Using the eFACE Facial Grading System

Importance: Assessing facial function using high-quality photographs would improve long-term and objective outcomes tracking in facial palsy, facilitate automated facial grading using artificial intelligence algorithms, and allow for remote follow up. Objective: To determine agreement between in-person and photographic electronic facial function scale (eFACE) assessments, and evaluate inter-rater reliability of photographic eFACE evaluation. Design, Setting, and Participants: Retrospective review of eFACE scores from in-person interviews and standardized photographs using the Massachusetts Eye and Ear (MEE) Standard Facial Palsy Dataset. Main Outcomes and Measures: eFACE total scores and subset scores determined by two experienced facial reanimation surgeons in person and from photographs. Results: Intraclass correlation coefficients for eFACE scores were 0.96 (95% CI 0.94 to 0.97) for total scores, 0.99 (95% CI 0.989 to 0.995) for static scores, 0.82 (95% CI 0.74 to 0.88) for dynamic scores, and 0.98 (95% CI 0.97 to 0.99) for synkinesis scores. Photographic and in-person scores had a mean difference of -0.64 (95% CI -2.05 to 0.77; p = 0.37) for total score, -1.58 (95% CI -4.22 to 1.05; p = 0.24) for the static subset, 0.14 (95% CI -1.70 to 1.97; p = 0.88) for the dynamic subset, and -1.11 (95% CI -3.09 to 0.86; p = 0.26) for the synkinetic subset. Bland-Altman analysis showed no trend for increasing differences in total score or subset scores. Conclusions: eFACE assessment obtained via photographs exhibits excellent inter-rater reliability and strong agreement with in-person assessment, demonstrating facial symmetry in facial palsy patients can be monitored using standardized frontal photographs.