Project Overview
Real-time eye blink tracking feature within a Flutter-based health app aimed at improving eye habits and detecting dry eye patterns. It combines video playback with front camera tracking and machine learning-based blink detection (using MediaPipe or TFLite) to monitor user eye behavior in real-time. The system accurately distinguishes full and partial blinks, calculates blink rate, and presents results in a user-friendly format—all while ensuring smooth performance on mobile devices. The feature delivers actionable insights for eye health and is optimized for scalability and future ML enhancements.
Client
Marcus
Industry
Health & Wellness
Services
Mobile App Development, Backend Development, UI/UX Design
Completed
August, 2025
Project Gallery
The Challenge
Monitoring eye behavior during screen usage requires real-time camera tracking, accurate blink detection, and smooth performance on mobile devices. The client needed a solution that could analyze blinking patterns while users watched video content and convert this data into meaningful health insights.
The main challenge was integrating machine learning–based eye tracking with video playback inside a Flutter mobile application without compromising performance or user experience.
Real-Time Camera & Video Processing
Running front-camera tracking while playing video simultaneously required careful performance optimization to prevent lag and frame drops.
Accurate Blink Detection
The system needed to reliably distinguish between full blinks and partial blinks using precise eyelid movement thresholds.
Mobile ML Integration
Existing machine learning logic needed to be adapted for mobile use through on-device processing or API-based architecture.
Our Approach
Custom Flutter Tracking Screen
Developed a dedicated Flutter interface that combines video playback with front-camera tracking in a single seamless experience.
Flexible ML Integration
Integrated MediaPipe / TFLite for on-device processing or connected with an existing Python ML backend through APIs.
Simultaneous Camera & Video Integration
Enabled real-time camera capture while the video player runs, ensuring both processes operate smoothly together.
Blink Pattern Analysis
Calculated key metrics such as blink rate (blinks per minute) and percentage of full blinks after each session.
Real-Time Blink Detection Logic
Implemented frame-by-frame eye analysis to detect blinking patterns and classify full and partial blinks.
User-Friendly Result Display
Designed a clean results interface that presents blink metrics and insights in an easy-to-understand format.
Performance Optimization
Optimized frame processing and state management to maintain smooth performance even on mid-range mobile devices.
Results & Impacts
Review
What Our Client Say
The thoughtful review from client upon project completion.
John Doe
Codetic