Detection of Stroke-Induced Facial Paralysis Using a Convolutional Neural Network
DOI:
https://doi.org/10.26821/IJSHRE.13.06.2025.130601%20Keywords:
Stroke-Induced Facial Paralysis, Convolutional Neural Network, Deep Learning in Healthcare, Stroke Diagnosis, Bell’s Palsy Diagnosis, Early Stroke Detection, Cost-Effective Medical Solutions, Facial Asymmetry DetectionAbstract
Stroke is one of the leading causes of mortality and disability. It can result in complications like facial paralysis which impair communication and emotional expression. Stroke cases are rising in India due to hypertension, diabetes, and cardiovascular diseases. There is an urgent need for affordable diagnostic tools, particularly in rural areas where healthcare access is limited. Diagnostic methods like clinical examinations and imaging systems, are either subjective, time-intensive, or expensive. This study focuses on addressing these challenges by developing a convolutional neural network (CNN) for detecting stroke-induced facial paralysis. The CNN model was trained on a dataset of 4,224 labeled images which were categorized into six classes representing different facial regions affected by paralysis along with severity. The CNN model consisted of multiple convolutional layers for extracting features. Max-pooling layers were used for reducing dimensionality, and dropout layers were used to avoid overfitting. The trained model achieved a validation accuracy of 97.04% with a validation loss of 0.0556. Its confusion matrix demonstrated the model’s accuracy in classifying samples across various classes. A Streamlit-based web application was developed to allow users to upload or capture images of their face for detecting facial paralysis symptoms that indicate stroke. The app also recommends on the basis of severity of facial paralysis detected in the captured or uploaded image. The results highlight the developed model's practical utility in early stroke detection. This model reduces diagnostic costs and enables accessibility for people living in areas that are underserved. This model can empower individuals with a user-friendly solution for effective stroke management.
References
Pandian, Jeyaraj Durai, and Paulin Sudhan. "Stroke epidemiology and stroke care services in India." Journal of stroke 15.3 (2013): 128.
Hospitals, A. (n.d.). Apollo Hospitals. Apollo Hospitals. https://www.apollohospitals.com/diseases-and-conditions/bells-palsy-symptoms-causes-risk-factors-and-treatment/#
House-Brackmann Facial Paralysis Scale | Facial Paralysis Institute. (n.d.). https://facialparalysisinstitute.com/conditions/house-brackmann-grading-system/
Song, Anping, et al. "Neurologist standard classification of facial nerve paralysis with deep neural networks." Future Internet 10.11 (2018): 111.
Gaber, A., Taher, M.F., Wahed, M.A. et al. Classification of facial paralysis based on machine learning techniques. BioMed Eng OnLine 21, 65 (2022). https://doi.org/10.1186/s12938-022-01036-0
Parra-Dominguez, Gemma S., Raul E. Sanchez-Yanez, and Carlos H. Garcia-Capulin. "Facial paralysis detection on images using key point analysis." Applied Sciences 11.5 (2021): 2435.
Chiesa-Estomba, Carlos Miguel, et al. "Machine learning models for predicting facial nerve palsy in parotid gland surgery for benign tumors." journal of surgical research 262 (2021): 57-64.
Guarin, Diego L., et al. "A machine learning approach for automated facial measurements in facial palsy." JAMA facial plastic surgery 20.4 (2018): 335-337.
Ali, Ahmad., Muhammad, Usama., Yasir, Niaz, Khan. "Early stroke disease prediction with facial features using convolutional neural network model." IAES International Journal of Artificial Intelligence, 13 (2024).:933-933. doi: 10.11591/ijai.v13.i1.pp933-940
Rao, Akshita A., Jacqueline J. Greene, and Todd P. Coleman. "Machine Learning Methods to Track Dynamic Facial Function in Facial Palsy." medRxiv (2024): 2024-03.
Ponnusamy, Muruganantham. "Detection of Facial Paralysis in Human Beings Using Machine Learning Clustering." 2023 3rd International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME). IEEE, 2023.
Gogu, Sridhar Reddy, and Shailesh R. Sathe. "autoFPR: An Efficient Automatic Approach for Facial Paralysis Recognition Using Facial Features." International Journal on Artificial Intelligence Tools 32.02 (2023): 2340005.
Bai, Lisha, Yujian Li, and Haojie Chang. "Classification of Facial Paralysis Based on Machine Vision Techniques." Academic Journal of Computing & Information Science 7.6: 95-99.
Ge, Xuri, et al. "Automatic facial paralysis estimation with facial action units." arXiv preprint arXiv:2203.01800 (2022).
Ali, Wasif, et al. "A transfer learning approach for facial paralysis severity detection." IEEE Access (2023).
Liu, Yuxi, et al. "Automatic assessment of facial paralysis based on facial landmarks." 2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML). IEEE, 2021.
Boochoon, Kieran, et al. "Deep learning for the assessment of facial nerve palsy: opportunities and challenges." Facial Plastic Surgery 39.05 (2023): 508-511.
