Crowd Control


Person of Interest TV series

What it does

Detect, classify, and find people in a crowd and analyze their paths and behaviors. The system connects to the cameras in all the Hajj area from the Holy Mosque to Mina, and provide the system admins with:

  • insights about the crowd from busy areas and crowd movement speed.
  • Detecting through facial recognition blacklisted people and notify security on their locations automatically.
  • Detecting through facial recognition missing people and notify organizers about their locations.

How we built it

The system consists of these main parts:

  • Camera Proxy: connect to cameras and send video feed to the backend.
  • Image Processing and learning module: Process video feed and detect information then send it to the backend.
  • Backend: Handle communication between modules and connect to the database with REST APIs exposed.
  • Dashboard: The control panel to add persons and see reports.
  • Mobile App: Recieve notifications of detected people

Challenges we ran into

Face detection accuracy and speed.

Accomplishments that we're proud of

Built Image detection module, backend, dashboard, and an app in less than 2 days.

What we learned

  • Time management
  • Tasks assigning
  • Importance of teamwork
  • Brainstorming
  • Problem-solving skills

What's next for Kashef

  • Improve the face detection algorithm.
  • Increase the scalability of the system.
  • Improve the AI.
  • Develop data insights.
Share this project: