Online proctoring system

Simple web application analyzes user actions for video proctoring systems. It automatically tracks the user’s gaze direction and hand position. The system detects the user’s departure and the presence of a second person. The presence of a voice is also detects.

Build and run the application locally

streamlit run main.py

Build and run the application in Docker

# Build Docker image
docker build . -t streamlit
# Run Docker container
docker run -p 8501:8501 streamlit

If the Docker container is running locally, the application will be available at http://localhost:8501.

Architecture

Examples

User leaving the frame

The direction of view

Recording of the detected second person in the frame onto the video file