Live-Learn Assist: AR Accessibility for Smart Learning

Live-Learn Assist – AI-powered AR accessibility app for smarter learning.
Clean and simple app interface designed for accessibility-first learning.
Select images directly from gallery for instant text detection and learning support.
Real-time text detection with LIVE Accessibility Mode.
Dyslexia-friendly reading mode highlighting complex words for easier understanding.

Inspiration

A lot of students have problems with sloppy handwritten notes, with complicated educational terms or visual problems, such as dyslexia. I desired to create an accessibility solution, which is real-time and easy to learn by changing hard-to-read text into readable and clean content immediately.

What it does

Live-Learn Assist AI-based AR application is an app that recognizes handwritten or tricky text on the screen based on a YOLOv8 computer vision model and superimposes the screen with clear and understandable text. It aids students to comprehend some educational contents in real time.

How I built it

I applied a YOLOv8 model that was trained on a dataset of handwritten text to identify regions of text. The app takes in live camera feeds, generates bounding boxes and extracts text with the aid of OCR. The system is programmed with RTX 3050 device to be graphically accelerated to ensure rapid real-time performance.

Challenges I ran into

Real time detection demanded a compromise between performance and accuracy. One of the primary technical issues was to optimize the model to run on GPUs and provide a seamless AR overlay as the camera frames are processed.

What I learned

This project helped me to enhance my computer vision, real-time AI processing, design (accessibility-oriented), and mobile AI integration skills.

What's next

The enhancements in the future will involve the implementation of simplified AI-generated explanations, multi-language features, and advanced accessibility features to visually impaired learners.