A drowsiness detection application created using mediapipe, streamlit and streamlit-webrtc

No description available

Driver Drowsiness Detection System A real-time Driver Drowsiness Detection System built using Python, OpenCV, MediaPipe, and Streamlit

Real-time driver drowsiness detection system using MediaPipe face landmarks, EAR/MAR analysis, Streamlit interface, and audio alerts.

Smart AI-based vehicle safety system using Streamlit for drowsiness detection and damage assessment, integrating MediaPipe and pretrained CNN models.

Real-time drowsiness monitoring app using MediaPipe, Streamlit, and Python. Alerts the driver with an alarm when signs of drowsiness (based on EAR & MAR) are detected. Includes UI controls and live detection graph.

Real-time AI-based driver drowsiness detection system using MediaPipe Face Mesh and Eye Aspect Ratio (EAR) with multi-level fatigue alerts and animated eye monitoring built using Streamlit.

A real-time ADAS system integrating 8 safety modules including lane detection, FCW, and drowsiness monitoring using PyTorch, YOLO, UNet, and MediaPipe, trained on ~160 GB datasets with a Streamlit UI.

Driver Monitoring System using EAR/MAR detection to identify drowsiness and yawning in real time. Uses OpenCV, MediaPipe, and Streamlit to track eye aspect ratio, mouth activity, and head pose. Includes live webcam feed, event logs, and alert system to detect unsafe driver behavior.

Real-time web-based drowsiness detection system using MediaPipe FaceMesh and Eye Aspect Ratio (EAR). Runs directly in the browser via Streamlit and WebRTC, allowing anyone to test eye-closure alerts with their webcam. Optimized for desktop, mobile, and Raspberry Pi, with live EAR charts, FPS metrics, and visual “Drowsy” warnings.

This project designs and develops a modular Driver Behavior Scoring System that processes driving telemetry data, performs real-time face and eye detection (for drowsiness/distraction) using OpenCV and MediaPipe, and uses machine learning to score driver safety and risk levels. All results are presented in an interactive Streamlit web application.