Inspiration
Statistics stated that most robbing happened in lift of US. Furthermore, most robbing is with knife. Therefore, needing a video detector that send alert notification reducing crime rate.
What it does
Web Application that used to receive notification when there is a suspicion of crime: The app receives videos (e.g. CCTV in real-time). In back-end, the app uses the computer vision technology to distinguish objects (e.g. knife and person) in the video such that it can evaluate the possibility of a crime occurring in real time. When there is a threat, the user (e.g. guards) will receive a notification from the app to decide whether he should take an action (e.g. ignore it / call the police).
How we built it
Mainly in Python.
Challenges we ran into
Obviously, debugging. We had spent too much time on some silly mistakes. Time was insufficient to add more features such as audio detection in our app.
Accomplishments that we're proud of
The completion of our app in the Hackathon. (Most team members in this group join Hackathon for the first time.) We enjoy the process of the idea & app design brainstorm, code writing and challenges overcoming.
What we learned
Team Collaboration and Communication. Time Management. Take challenge. Risk taker.
What's next for Elesafer
Audio Detection: When one is yelling and shouting loudly, it is also an alert. Together with the existing object detection functions, the receipt of double alerts means that it is most likely that there is a crime. This makes the app become more reliable in crime detection. Any follow-up actions taken for the developers when the alert is false. Think about how to raise the precision rate, for instance, in different training algorithms. Identifying more objects such as guns.
Log in or sign up for Devpost to join the conversation.