A Novel Approach to Detect Face Mask using CNN

Abstract

Face detection and recognition will be considered as one of the most intriguing modalities for biometric models. Those researches are mainly required for ensuring security in a most sensitive area. This research paper has proposed a very fast image pre-processing with the mask in the center over the faces. For this system, features extraction and Convolutional Neural Network are used for classification and detection of a masked person. This research work will be carried out in three levels: preprocessing the images, cropping the images and classification of the images. This helps to detect whether the face is masked or not. A webcam or CCTV camera surveillance will record all the time and it checks whether the person is wearing a mask, if the person doesn't wear a mask then the system gives a security alert.

Face Mask Recognition Menggunakan Model CNN (Convolutional Neural Network) Berbasis Python dan OpenCV

“…213 negara terkena dampaknya coronavirus, termasuk semua negara berkembang seperti AS, Inggris, Rusia, Cina, Jepang, Italia, dll. [10] Penyakit virus corona 19 atau COVID-19 adalah penyakit pernapasan yang menyebabkan parah pneumonia pada orang yang terinfeksi dan tertular penyakit ini adalah melalui kontak langsung orang-ke-orang dengan dihasilkan tetesan pernapasan, tetesan air liur atau debit dari hidung ketika orang yang terinfeksi batuk atau bersin atau melalui menghirup virus jika Anda berada dalam jangkauan yang terinfeksi orang [11] Mengenakan masker wajah telah direkomendasikan untuk mengurangi penularan virus dan menyelamatkan nyawa sejak wabah awal. [12] Selama periode waktu bencana ini setiap orang harus menaikkan kesadaran dan secara alami harus melakukan beberapa kegiatan sendiri.…”

Detecção de uso de máscaras em ambientes fechados com MobileNetV2 e Single Shot Multibox Detector

“…O trabalho desenvolvido por Islam et al (2020) introduz uma abordagem para um sistema de detecc ¸ão e alerta do uso de máscaras de protec ¸ão utilizando redes convolucionais. Nesse sistema é adotado uma arquitetura simples de rede convolucional com a saída sendo dois valores representando respectivamente as probabilidades das classes with mask e without mask.…”

“…Ao passo que a utilizac ¸ão de redes convolucionais é sem dúvidas eficiente, o trabalho possui dois grandes problemas, a não utilizac ¸ão de um backbone (uma rede de classificac ¸ão servindo como feature extractor), o que acarreta na necessidade de utilizar grandes volumes de dados e/ou um longo período de treinamento, e a detecc ¸ão apenas de uma pessoa por imagem. [Islam et al 2020] Jiang e Fan (2020) propõem o que consideram a primeira arquitetura de alta performance para um detector de faces no contexto de uso de máscaras faciais. O trabalho apresentado descreve caracteríscas não vistas em trabalhos anteriores como distinguir máscaras usadas incorretamente e corretamente.…”

Face mask recogniser using image processing and computer vision approach

“…This can help to spread awareness and educate people the correct way to wear a mask. This implementation can help frontline workers focus on eradication of the virus [8] . This face mask recogniser is a necessity and as a face mask is our shield against the virus, developing this model was essential and presently with the scare of the new variants, this finds high application value which motivated the idea for this study.…”