Every person in the pandemic has to put on a
mask to stop the CORONA virus from spreading. In
these difficult COVID-19 times, developing a model that
detects people with and without masks in real-time is
critical as a simple precautionary measure to prevent
virus spread. If used correctly, this machine learning
technique can help frontline warriors simplify their
work while also saving their lives. Tensor Flow, Keras,
and OpenCV are used in the development of a
convolutional neural network (CNN) model, which helps
the algorithm make the most accurate predictions. The
Java-script API facilitates webcam access for face mask
detection in real time. The first stage, known as preprocessing, consists of "grayscale conversion" of an
RGB image and "image resizing and normalisation" to
prevent inaccurate predictions. As the output layer of
the proposed CNN architecture has two neurons with
Soft max activation to classify the same, the suggested
CNN then distinguishes between facial characteristics
with and without masks. The suggested design has a
validation accuracy of 96%. If anyone in the video a
green rectangle is drawn around the appearance of a
person using a mask, while a red rectangle with the
words "NO MASK" is drawn around the face of stream.
Keywords :
CNN Model, Java Script, APIs, Deep Learning, Tensor Flow, Keras, Open CV, Pandas, Performance, System Architecture, Adaptive Models.