To enforce social distancing, it is important to monitor suspicious scenarios like crowd gathering or tracking etc. Usage of AI has surpassed traditional methods and can help automate this pattern recognition. The objective of this problem statement is to find different patterns on input sources like camera, satellite, social platforms which can help identify suspicious activities to stop or track covid-19 spread.
Apollo’s Eye is a smart web application deployed using Streamlit framework (executed by a launcher) where the end user can upload a particular video or fetch data from the live surveillance cameras located at hospitals or public places in order to identify instances where social distancing norms are been violated -- tracking face masks and crowd formation using proximity detection. The user receives a visual alert, with the help of which they can enforce action and disperse the crowd and enforce action. Our application has a high mask and crowd distinction rate which is higher and more effective than the existing state-of-art systems.
- We can enhance our model to detect any form of obstruction on railway tracks. In 2020, a huge number of migrant laborers lost their lives after sleeping on railroads. Furthermore, the death of animals after getting hit by a train is a major issue that has been a factor of concern for many years. We wish to aid these problems using our solution to create an alert to local authorities/police using camera systems installed at appropriate locations; so that they can clear off the track before any lives are lost.
- Our solution can be formulated and deployed in the form of a smartphone app architecture, through which any person using it will be able to report problematic crowd activities like street brawls or strikes using live feed/images so that the police or patrol can take immediate actions to disperse them using location.