Detection of Face Mask Wearing Condition for COVID-19 using Mask R-CNN

Abstract

Due to the COVID-19 pandemic, which has affected the whole world, countries have made it mandatory for people to wear face masks. Because wearing a mask is considered one of the most effective methods to reduce the risk of transmission of the virus. However, it is difficult to manually check whether people are wearing masks. It is aimed to develop a model that detects all kinds of face masks in crowded environments using a deep neural network in this study. Mask R-CNN, which is one of the deep learning algorithms and used for object detection was used to detect and classify people’s mask states. The proposed deep learning model was trained and tested with k-fold cross-validation using a dataset of 853 images containing three classes (with mask, without a mask, incorrect use of mask). ResNet101 backbone was chosen as the backbone architecture and transfer learning was performed using the COCO model. The proposed Mask R-CNN model achieves an mAP of 83%, an mAR of 90%, and an F1 score of 86%. These results reveal that the proposed model is successful in mask detection.

Keywords

• Covid-19
• Deep learning
• Mask R-CNN
• Mask detection

Mask R-CNN kullanarak COVID-19 için Yüz Maskesi Takma Durumunun Tespiti

Abstract

Tüm dünyayı etkisi altına alan COVID-19 salgını nedeniyle ülkeler insanların yüz maskesi takmasını zorunlu hale getirdi. Çünkü maske takmak virüsün bulaşma riskini azaltmak için en etkili yöntemlerden biri olarak kabul edilmektedir. Ancak insanların maske takıp takmadığını manuel olarak kontrol etmek zordur. Bu çalışmada derin bir sinir ağı kullanılarak kalabalık ortamlarda her türlü yüz maskesini algılayan bir modelin geliştirilmesi amaçlanmıştır. Derin öğrenme algoritmalarından biri olan ve nesne tespiti için kullanılan Mask R-CNN, insanların maske durumlarını tespit etmek ve sınıflandırmak için kullanıldı. Önerilen derin öğrenme modeli, üç sınıf (maskeli, maskesiz, yanlış maske kullanımı) içeren 853 görüntüden oluşan bir veri seti kullanılarak k-kat çapraz doğrulama ile eğitildi ve test edildi. Omurga mimarisi olarak ResNet101 seçildi ve COCO modeli kullanılarak transfer öğrenmesi gerçekleştirildi. Önerilen Mask R-CNN modeli, %83'lük bir mAP, %90'lık bir mAR ve %86'lık bir F1 puanına ulaşmıştır. Bu sonuçlar önerilen modelin maske tespitinde başarılı olduğunu ortaya koymaktadır.

Keywords

• Covid-19
• Derin öğrenme
• Mask R-CNN
• Maske tespiti

References

1. Sardogan, M., Tuncer, A., and Ozen, Y., Plant Leaf Disease Detection and Classification Based on CNN with LVQ Algorithm, In 2018 3rd International Conference on Computer Science and Engineering (UBMK), IEEE, 382-385, (2018).
2. Orman, A., Köse, U., and Yiğit, T., Açıklanabilir Evrişimsel Sinir Ağları ile Beyin Tümörü Tespiti, El-Cezeri Fen ve Mühendislik Dergisi, 2021, 8(3): 1323-1337.
3. Sardogan, M., Özen, Y., and Tuncer, A., Detection of Apple Leaf Diseases using Faster R-CNN, Düzce Üniversitesi Bilim ve Teknoloji Dergisi, 2020, 8(1): 1110-1117.
4. Girshick, R., Donahue, J., Darrell, T., and Malik, J., Region-Based Convolutional Networks for Accurate Object Detection and Segmentation, IEEE Trans. Pattern Anal. Mach. Intell., 2015, 38(1): 142-158.
5. Ren, S., He, K., Girshick, R., and Sun, J., Faster R-CNN: Towards Realtime Object Detection with Region Proposal Networks, IEEE Trans. Pattern Anal. Mach. Intell., 2017, 39(6), 1137- 1149.
6. Redmon, J., Farhadi, A., YOLOv3: An Incremental Improvement, 2018, arXiv preprint arXiv:1804.02767.
7. He, K., Gkioxari, G., Dollár, P., and Girshick, R, Mask R_CNN, Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2961-2969, (2017).
8. Amin, P. N., Moghe, S. S., Prabhakar, S. N., and Nehete, C. M., Deep Learning Based Face Mask Detection and Crowd Counting, In 2021 6th International Conference for Convergence in Technology (I2CT), IEEE, 1-5, (2021).

9. Bhuiyan, M. R., Khushbu, S. A., and Islam, M. S., A Deep Learning Based Assistive System to Classify Covid-19 Face Mask for Human Safety with YOLOv3”, In 2020 11th International Conference on Computing, Communication and Networking Technologies (ICCCNT), IEEE, 1-5, (2020).
10. Liu, R., and Ren, Z., Application of Yolo on Mask Detection Task, In 2021 IEEE 13th International Conference on Computer Research and Development (ICCRD), IEEE, 130-136, (2021).
11. Susanto, S., Putra, F. A., Analia, R., and Suciningtyas, I. K. L. N., The Face Mask Detection for Preventing the Spread of COVID-19 at Politeknik Negeri Batam, In 2020 3rd International Conference on Applied Engineering (ICAE), IEEE, 1-5, (2020).
12. Abbasi, S., Abdi, H., and Ahmadi, A., A Face-Mask Detection Approach based on YOLO Applied for a New Collected Dataset, In 2021 26th International Computer Conference, Computer Society of Iran (CSICC), IEEE, 1-6, (2021).
13. Gawde, B. B., A Fast, Automatic Risk Detector for COVID-19, In 2020 IEEE Pune Section International Conference (PuneCon), IEEE, 146-151, (2020).
14. Singh, J., and Shekhar, S., Road damage detection and classification in smartphone captured images using mask r-cnn, arXiv preprint arXiv:1811.04535, (2018).
15. Cakiroglu, O., Ozer, C., and Gunsel, B., Design of a deep face detector by mask r-cnn, In 2019 27th Signal Processing and Communications Applications Conference (SIU), IEEE, 1-4, (2019).
16. Bayram, F., Derin öğrenme tabanlı otomatik plaka tanıma, Politeknik Dergisi, 2020, 23(4): 955-960.
17. Lin, T.Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollar, P., and Zitnick, C. L., Microsoft coco: Common objects in context, European conference on computer vision. Springer, Cham, 740-755, (2014).
18. He, K., Zhang, X., Ren, S., Sun, J., Deep residual learning for image recognition, Proceedings of the IEEE conference on computer vision and pattern recognition, 770-778, (2016)
19. Lin, K., Zhao, H., Lv, J., Li, C., Liu, X., Chen, R., and Zhao, R., Face detection and segmentation based on improved mask r-cnn, Discrete dynamics in nature and society, (2020).
20. Mask Dataset. [Online]. Available: https://www.kaggle.com/andrewmvd/face-mask-detection
21. Chang, Y. Y., Li, P. C., Chang, R. F., Yao, C. D., Chen, Y. Y., Chang, W. Y., and Yen, H. H., “Deep learning- based endoscopic anatomy classification: an accelerated approach for data preparation and model validation”, Surgical Endoscopy, 2021, 1-11.