🧠 NeuroCursor Eye Controlled Cursor using Python
NeuroCursor is an AI-based eye tracking system that allows users to control their computer cursor using eye movements and blink gestures. It uses computer vision and facial landmark detection to track the user's eyes in real-time through a webcam. The system converts eye movement into mouse movement, making the computer accessible without physical mouse interaction. This project was developed as part of a hackathon innovation focusing on human-computer interaction and accessibility technology.
🚀 Features Eye Controlled Cursor The cursor moves based on real-time eye movement detection captured from the webcam.
Blink Gesture Controls.
Gesture Action Double Blink Left Click Triple Blink Double Click Look Down Scroll Down Look Up Scroll Up
🖥 Real-Time Computer Vision Uses facial landmarks detection to accurately track eye positions. Fast & Lightweight Runs smoothly using Python libraries without requiring heavy AI models. Accessibility Helps users who have difficulty using traditional mouse input devices.
Technologies Used Python Computer Vision Facial Landmark Detection Webcam Input Processing
📦 Libraries Used Install the following Python libraries before running the project: opencv-python mediapipe pyautogui numpy time
You can install them using: "pip install opencv-python mediapipe pyautogui numpy"
Note: time is a built-in Python module and does not require installation.
📂 Project Structure NeuroCursor/ │ ├── main.py # Main program file ├── README.md # Project documentation
The webcam captures the user's face. MediaPipe Face Mesh detects facial landmarks. Eye landmarks are extracted. Eye movement is analyzed to determine cursor direction. Blink patterns trigger different mouse actions. Cursor movement and clicks are performed using PyAutoGUI.
📸 System Requirements Python 3.8+ Webcam Windows / Mac / Linux Minimum 4GB RAM
💡 Future Improvements Right click gesture Cursor speed calibration GUI interface Better blink detection accuracy Head movement support
👨💻 Authors Developed by:
Team Codzen Leaded by Farhaad Ahmad
📜 License This project is open-source and available under the MIT License.