FocusPal

Inspiration

Our inspiration for FocusPal came from witnessing firsthand the challenges faced by individuals with ADHD in educational settings. Traditional learning environments often fail to adapt to different attention patterns, leaving many students feeling frustrated and underserved. We were particularly moved by research showing that personalized learning approaches can significantly improve outcomes for neurodivergent learners.

The increasing prevalence of remote learning environments during recent years highlighted how technology could either help or hinder focus, depending on its implementation. We saw an opportunity to create a solution that uses the very screens that can sometimes be distracting as tools for improving focus and learning outcomes.

Additionally, many team members have personal connections to ADHD, either through their family members and friends. We recognized the need for a tool that doesn't just accommodate but actively supports different learning styles and attention patterns.

What it does

FocusPal is an adaptive learning platform that uses computer vision to measure a user's focus levels in real-time and tailors educational content accordingly. The application:

Allows Users to upload study material (eg- Lecture Slides)
Detects attention levels by analyzing facial expressions, eye movements, and head angles using a webcam
Dynamically adjusts content format and difficulty based on the user's current focus state eg- Videos instead of text when focus dips a lot.
Introduces "micro-interventions" when focus wanes, such as brief interactive exercises, visual changes, or well-timed breaks
Provides users with insights about their personal focus patterns and learning preferences
Offers content in various modalities (text, video, quiz, interactive mini games, flash cards) and automatically switches between them based on effectiveness
Gamifies the learning experience with focus-based rewards and achievements
Creates a personalized learning profile that evolves as the system learns what techniques work best for each user

The platform also includes an anonymized research component, allowing educators and researchers to study attention patterns across different subjects and learning approaches while maintaining strict privacy standards.

How we built it

FocusPal was built using a combination of cutting-edge technologies:

Frontend: Next.js with Tailwind CSS for a clean, distraction-minimized UI
Backend: Flask using Python, providing the API for content delivery and user management. We also established web sockets for the live streaming of video and inference of the Focus levels.
Computer Vision: Integrated MTCNN for face detection and attention metrics
Gemini: Used Gemini to generate interactive content sometimes in the form of React Code that we would dynamically render on the frontend.
ReVid API: To generate reels/shorts about the concept at hand.
Authentication: Auth0 for secure user management and role-based access control
Database: MongoDB for storing user preferences, learning progress, and anonymized attention data

Challenges we ran into

Developing FocusPal presented several significant challenges:

Ethical Data Handling: Balancing the need for effective attention monitoring with privacy concerns was complex. We spent considerable time designing a system that could be effective without being invasive.
Attention Classification Accuracy: Determining a user's focus state based on a variety of visual cues proved difficult, especially across different lighting conditions. We had to refine our models extensively to improve reliability.
Content Adaptation Logic: Creating an algorithm that could make smart decisions about when and how to adapt content required numerous iterations. Early versions were too reactive, creating a distracting experience rather than a supportive one.
Technical Integration: Combining live computer vision processing with smooth content delivery using Web Sockets was challenging to implement effectively.
User Experience Design: Designing an interface that supports focus rather than detracts from it required careful consideration of every element.

Accomplishments that we're proud of

Despite the challenges, we achieved several outcomes that make us particularly proud:

Effective Focus Detection: Our system can detect changes in attention with impressive accuracy, even with standard webcams in varied environments.
Seamless Content Adaptation: The platform transitions between different content formats and difficulty levels smoothly, without disrupting the learning experience.
Technical Performance: We achieved real-time attention analysis and content adaptation with minimal latency, even on modest hardware.

What we learned

This project provided numerous learning opportunities for our team:

Technical Skills: We deepened our expertise in computer vision, network engineering, and full-stack development. Integrating these diverse technologies gave us valuable experience in building complex, interactive systems.
Human-Centered Design: We gained insights into designing for neurodivergent users and learned how to effectively engage with this audience.
Ethical AI Development: We developed a stronger understanding of privacy considerations and ethical implications when using AI to analyze human behavior.
Performance Optimization: We learned valuable techniques for optimizing resource-intensive applications to run smoothly on consumer hardware.
Adaptive Learning Principles: Researching and implementing adaptive learning techniques gave us deeper insights into educational psychology and effective learning strategies.

What's next for FocusPal

We have ambitious plans for FocusPal's future:

Enhanced Personalization: Refining our machine learning models to better predict optimal learning strategies for each user based on historical data.
Mobile Application: Creating a mobile version to support learning on smartphones and tablets with appropriate adaptations for smaller screens.
Educator Dashboard: Developing tools for teachers and tutors to monitor student progress and customize learning experiences.
API for Developers: Opening our platform to third-party developers who want to create specialized learning modules or integrate with existing educational software.
Research Partnerships: Establishing formal collaborations with educational institutions to study the effectiveness of adaptive learning techniques for ADHD.
Further Gamification: Gamifying the learning process can make learning fun even for those who have additional hurdles.

Our ultimate vision is to make FocusPal a comprehensive learning companion that adapts not just to attention differences but to the full spectrum of learning preferences and needs, making education more accessible and effective for everyone.