Face Mask Detection and Comparison between Different Image Classification Algorithms
Author Name : Jairaj Mahadev, Janani Iyer, Sarvesh Moraskar, Manoj Inbarajan, Rizwana S
The situation report of the World Health Organization presented that coronavirus disease has globally infected over 193 million people and caused over 4,000,000 deaths. At the moment, it is recommended that people should wear face masks if they have respiratory symptoms, or they are taking care of the people with symptoms. We have seen people who refuse to wear masks in public as they say its uncomfortable or it causes problems in breathing which is totally false. Many public service providers require customers to use the service only if they wear masks. This has proven to cause many issues. Also, it is uncertain that this virus is going to go away soon. Therefore, face mask detection has become a crucial computer vision task to help the global society. Themain motive of this project was to create a surveillance system that could detect if a person was not wearing a mask. A convolution neural network was used for the classification. The MobileNetV2 architecture was used as it is specifically designed for mobile and low powered devices. The dataset used was from Kaggle and was added on to using images from Google images and other sources by us. This program used various technologies like machine learning and computer vision. Our system was designed keeping low end devices in mind and therefore can be implemented successfully by any organization large or small.
Keywords:Coronavirus, Face Mask Detection, Convolution Neural Network, MobileNetV2 Architecture.