Dyslexi

Lexishift: Making World Easy for Dyslexic Readers

The Inspiration Dyslexia affects approximately 15-20% of the population, yet most digital content isn't formatted with their needs in mind. I witnessed firsthand how my cousin struggled with reading academic PDFs during college, often spending hours reformatting documents manually. This personal connection inspired me to create Lexishift, a tool that could help millions of people access written content more easily. What We Built Lexishift is a web-based PDF converter specifically designed for readers with dyslexia. The platform offers: Intelligent PDF processing with OCR technology Customizable text formatting options including font type, size, spacing, and color Machine learning-powered formatting suggestions Community features connecting users with specialists and support groups Integration with mental health professionals and NGOs

The Development Journey Technology Stack Frontend:

React.js
Tailwind CSS
PDF.js for document rendering

Backend:

Python/Flask
Tesseract OCR
TensorFlow for ML models
PostgreSQL database

Key Features Implementation

PDF Processing Pipeline The heart of Lexishift is its document processing system. We implemented OCR using Tesseract, which proved challenging with complex layouts but essential for accessibility. ML-Powered Formatting We trained our model on user preference data to suggest optimal formatting combinations. This feature evolved from user feedback showing that manual adjustments were time-consuming. Professional Network Integration We created a verified registration system for mental health specialists and NGOs, ensuring users can connect with legitimate support resources.

Challenges and Solutions

Technical Challenges

OCR Accuracy Challenge: Initial OCR results were inconsistent with complex layouts Solution: Implemented pre-processing steps to improve document quality and developed a custom post-processing algorithm

Performance Optimization Challenge: Processing large PDFs caused significant delays Solution: Added document chunking and parallel processing capabilities

User Experience Challenges

Accessibility Challenge: Making the interface itself dyslexia-friendly Solution: Collaborated with dyslexic users to design an intuitive, accessible UI

Community Building Challenge: Creating a safe, supportive environment Solution: Implemented moderation tools and verification systems for professional users

Learning Outcomes

This project taught me invaluable lessons about: The importance of inclusive design in technology Real-world application of ML in accessibility tools Building communities around technological solutions The intersection of mental health and software design

Future Developments We're currently working on: Mobile app development Integration with popular document management systems Expanded language support Enhanced community features