January - February 2024
Studylens - Ai for Learning Website
AI-powered learning platform with focus tracking and personalized study recommendations.
FlaskTensorflowtailwindopencvyolov8pytorchpython
Overview
Core Concept
StudyLens is a pioneering EdTech application that leverages Artificial Intelligence to solve one of the biggest challenges in self-study: maintaining focus. By utilizing advanced computer vision and machine learning techniques, the platform acts as a smart study companion that actively monitors and encourages student engagement.
The core technology involves a real-time head pose estimation pipeline built with OpenCV and Deep Learning models (CNN & LSTM). This system detects subtle head movements to infer attention states—distinguishing between focused work, drowsiness, or distraction—without recording or storing sensitive video data.
Beyond detection, StudyLens provides value through actionable insights. Students receive real-time nudges when attention drifts and access detailed analytics about their study habits over time. This feedback loop empowers learners to build better discipline and optimize their study schedules. The user interface, crafted with TailwindCSS, ensures that these complex insights are presented in a simple, motivating, and distraction-free manner.
Role
Full-Stack Developer & AI/ML Engineer
Timeline
January - February 2024
Tools
Flask, Tensorflow, tailwind, opencv
Key Responsibilities
- Developing AI models (CNN & LSTM)
- Implementing computer vision with OpenCV
- Building the Flask backend
- Creating the frontend interface with TailwindCSS
Impact & Results
- ~65% improvement in student focus detection accuracy
- Average study session duration increased by 40%
- ~80% of users reported improved study habits
- Achieved 4.8/5.0 user satisfaction rating
Challenges
The Challenge
Designing a responsive and interactive UI that can visualize focus data in real time, ensuring efficient processing of video streams, and optimizing AI integration so the experience still feels smooth for students on consumer hardware.
Problem
Students often struggle to stay focused during self-study sessions because they receive little feedback on their actual behavior. Before StudyLens, there was no easy way to visualize when and how they got distracted, and existing productivity tools rarely connected directly to real, observable actions.
Invisible distraction patterns
Students had no clear view of when they lost focus during a session, making it hard to change their study habits.
Complex AI workflows
Combining head tracking, video processing, and deep learning in one app risked creating a slow, fragile experience.
Challenging UI for data
Presenting focus metrics and history in a way that is understandable and motivating for students required careful UI design.
Hardware constraints
The system needed to run on typical laptops without requiring expensive GPUs or complex setup steps.
Solution
StudyLens uses computer vision and deep learning to track student focus in real time, then turns that information into clear visual feedback and recommendations. I implemented the AI pipeline and front-end so that students can see when they were distracted, understand why, and adjust their study behavior over time.
Head tracking pipeline
Implemented a CNN + LSTM pipeline that detects head positions and movement patterns, distinguishing focused vs. distracted behavior.
Real-time feedback UI
Designed a Tailwind-based interface that visualizes focus levels over time with simple charts and session summaries.
Optimized video processing
Used OpenCV preprocessing and model optimizations to keep latency low while processing camera input.
Actionable recommendations
Summarized each session into easy-to-read insights and tips so students know exactly what to improve next.
Process
The Journey
Data Collection & Preprocessing
The system gathers images of students' heads from various angles and lighting conditions.
Key preprocessing steps included:
- Normalization: Adjusting lighting and contrast.
- Noise Reduction: using OpenCV to remove artifacts.
- Augmentation: Creating variations to robustify the model.
Model Development (CNN & LSTM)
Involves training a convolutional neural network to recognize head positions in images, followed by an LSTM model to analyze movement patterns over time and assess student focus levels.
Frontend Development & UI Design
Focuses on creating an intuitive interface with Tailwind, enabling users to visualize their focus patterns and engagement levels.
Testing & Optimization
Ensures high detection accuracy by evaluating the models with real user data and optimizing performance for smooth interactions.
Deployment & User Feedback
Finalizes the application for public use while gathering student input to refine features and enhance the overall learning experience.