A.A. Boiko1, M.V. Dembovskiy2

1, 2 Bauman Moscow State Technical University (Moscow, Russia)
 

Diseases associated with lesions of the facial nerve are quite common and can be caused by various reasons. Treatment of lesions of the facial nerve can be both surgical and conservative. In order to control the treatment process, various scales have been developed that assess the functioning of the facial nerve. Scales in most cases are ordinal and include from 6 to 100 gradations. The disadvantage of the developed scales is the element of subjectivity in assessing the degree of damage to the facial nerve. To eliminate this shortcoming, attempts are being made to automate such an assessment. A number of automation approaches rely on the use of facial landmarks coordinates. In this case, the symmetry indices of the upper and lower parts of the face are evaluated separately. Despite a significant number of works devoted to automation, there is not enough data in the literature on the ranges of values in which the symmetry indices can be found. In the present work, the values of the range of the facial symmetry index in the region of the forehead and mouth were obtained when performing eight tests for motor activity of the mimic muscles of the face. For this purpose, video images of the process of performing tests by 20 healthy subjects were recorded, of which 5 subjects were female and 15 were male. The age of the subjects ranged from 22 to 37 years. Registration was performed using a Logitech C922 Pro Stream Webcam in video recording mode. The distance from the video camera to the subject was from 50 to 70 cm, behind the subject there was a light uniform background. The resolution of the resulting video image was 1920 x 1080 pixels, the recording speed was 30 fps. The resulting video images were marked into intervals of performing individual tests, then automatic placement of facial landmarks on the images and calculation of the symmetry index in the forehead and mouth were performed. To calculate the symmetry index in the forehead area, the distances from the midpoint of the eyebrow to the midpoint of the eye on the right and left sides of the face were analyzed. To calculate the symmetry index in the mouth area, the distances from the outer corner of the mouth to the center of the mouth on the right and left sides of the face were analyzed. As a criterion for the "quality" of the indices, the minimum range of values for all subjects was considered. According to this criterion, the most promising is the use of the facial symmetry index in the forehead area when performing the “eyebrows raising” and “eyes squeezing” tests, as well as the symmetry index in the mouth area when performing the “smile”, “forced smile” and “lip stretching” (“lips as a tube”) tests.

Boiko A.A., Dembovskiy M.V. Experimental study of facial dynamical symmetry indices. Biomedicine Radioengineering. 2023. V. 26. № 2. P. 50–56. DOI: https://doi.org/ 10.18127/ j15604136-202302-08 (In Russian)

1. Shevcova E.A. K voprosu o diagnostike, lechenii i prognoze idiopaticheskoj dvuhstoronnej nevropatii licevogo nerva u detej i podrostkov. Vselennaya mozga. 2019. T. 1. № 3 (3). S. 41–43 (In Russian).
2. Bakaev A.A. Hirurgicheskoe lechenie intratemporal'nyh porazhenij licevogo nerva: Diss. … kand. med. nauk. M. 2022. C. 136 (In Russian).
3. Akulov M.A., Orlova O.R., Tabashnikova T.V., Karnauhov V.V., Orlova A.S. Porazhenie licevogo nerva pri nejrohirurgicheskih operaciyah: reabilitacionnyj potencial botulinoterapii. Voprosy nejrohirurgii im. N.N. Burdenko. 2018. T. 82. № 1. S. 111–118. https://doi.org/10.17116/neiro2018821111-118 (In Russian).
4. Song A. et al. Neurologist standard classification of facial nerve paralysis with deep neural networks. Future Internet. 2018. V. 10. № 11. P. 111.
5. Parra-Dominguez G.S., Garcia-Capulin C.H., Sanchez-Yanez R.E. Automatic Facial Palsy Diagnosis as a Classification Problem Using Regional Information Extracted from a Photograph. Diagnostics. 2022. V. 12. № 7. P. 1528.
6. Kim H.S. et al. A smartphone-based automatic diagnosis system for facial nerve palsy. Sensors. 2015. V. 15. № 10. P. 26756–26768.
7. Holze M. et al. Learning from EMG: semi-automated grading of facial nerve function. Journal of Clinical Monitoring and Computing. 2022. V. 36. № 5. P. 1509–1517.
8. Lavrova E.A., Samorodov A.V., Mordovskij A.V., Kudrin K.G., Polyakov P.A. Biotekhnicheskaya sistema avtomaticheskoj ocenki stepeni povrezhdeniya licevogo nerva. Biomedicinskaya radioelektronika. 2018. № 7. C. 4–6 (In Russian).
9. Dlib C++ Library. URL: http://dlib.net (data obrashcheniya: 22.02.2023).
10. Wu Y., Ji Q. Facial landmark detection: A literature survey. International Journal of Computer Vision. 2019. V. 127. P. 115–142.
11. Xia Y. et al. AFLFP: A Database With Annotated Facial Landmarks for Facial Palsy. IEEE Transactions on Computational Social Systems. 2022.
12. Guarin D.L. et al. Toward an automatic system for computer-aided assessment in facial palsy. Facial Plastic Surgery & Aesthetic Medicine. 2020. V. 22. № 1. P. 42–49.