How abdullahsajid007/eye-controlled-mouse-using-mediapipe-and-opencv works
Project Type and Purpose
• It’s a single‐script, real‐time computer‐vision application in Python.
• Purpose: capture webcam video, detect eye position and blinks, map movements to mouse control.
Main Architectural Components
a. Input Layer
• Physical Webcam (video stream)
b. Processing Layer (data‐flow pipeline)
1. Video Capture Module
– OpenCV captures frames from webcam.
2. FaceMesh Detector
– Mediapipe FaceMesh processes each frame → outputs facial landmarks (eye coordinates).
3. Landmark Processor
– Extracts relevant eye landmarks, computes normalized positions & blink metric.
4. Coordinate Mapper
– Maps normalized eye‐position to screen coordinates based on display resolution.
5. Blink Detector
– Computes vertical distance between eye landmarks → detects blink event.
c. Output Layer
• Mouse Control Module (PyAutoGUI)
– Receives mapped coordinates and blink events → moves cursor & issues click events.
d. Optional Visualization
• OpenCV window showing live feed + overlayed landmarks.
Relationships & Data Flow
[Webcam] → [Video Capture] → [FaceMesh Detector] → [Landmark Processor] →
branching to → [Coordinate Mapper] → [Mouse Control] ⇒ OS cursor
and → [Blink Detector] → [Mouse Control] ⇒ OS click
Architectural Patterns & Design Principles
• Pipeline pattern: discrete, sequential processing stages.
• Single‐process, event‐driven loop.
• Modular separation (even in one script): input, detection, mapping, output.
Technologies & Frameworks
• OpenCV: video I/O & visualization.
• Mediapipe: face landmarks detection.
• PyAutoGUI: OS‐level mouse control.
• Python: orchestration.
Required Elements Checklist
• If desired, note future-enhancement extension points (e.g., “Drag‐and‐Drop Gesture Detector”).