In recent trend in world wide Lockdowns due to COVID19 outbreak, as Face Mask is became mandatory for everyone while roaming outside. Masks play a crucial role in protecting the health of individuals against respiratory diseases, as is one of the few precautions available for COVID-19 in the absence of immunization.
Is it possible to create a model to detect people wearing masks, not wearing them.
An ML (deep learning) model which can analyzing whetaher a face is covered with mask or not.
- Deployed web app : https://praathapj-facemaskdetection-app-b5c2wm.streamlit.app/
Real world images obtained from kaggle: https://www.kaggle.com/datasets/omkargurav/face-mask-dataset
1100 images for both class, obtained 99% accuracy with ImageNet pretrained model.
- Evaluation metric is binary cross entropy.
- Obtain more data set with has images of 'mask not worn properly'.
- Also include images which has face covered without mask as 'No Mask'