Inspiration

Dyslexia is a learning disorder that affects how a person is able to read, write & spell. Individuals suffering from dyslexia face difficulty in recognizing and pronouncing words affecting their overall learning capabilities. Close to 5-10% of the population suffers from dyslexia affecting how they are able to interact with others. Currently, there is no product available that provides tailored services catering to differing severities of dyslexia.

What it does

The solution that we are proposing is an AI powered system aimed at providing personalised learning support as well as assistance to people suffering from dyslexia. Level of dyslexia differs from individual to individual and it is not a wise approach to treat everyone the same. We intend to first determine the severity of dyslexia and provide personalised learning paths. We have used standard dyslexic screeners used by psychologists to gauge the level of dyslexia and then generated a score. Then depending on the severity, we suggest learning exercises and use a gamification approach for the individual to improve their overall ability to learn and interact.

How we built it

Dyslexic Screening Tests - Generating a dyslexia score through a series of industry approved dyslexia tests utilizing AI and OCR technologies.

Dyslexia Severity Identification - Utilizing a Random Forest based ML model on the dyslexia score to identify dyslexia severity

Personalized Learning Path Mild - Reading Passages & Comprehensions Moderate - Memory Games Severe - Phonological Improvement Tools

Notes & Mind Map Generation Providing learning aids to an individual by DistilBERT based personalized notes and diagrammatic mind map generation.

Reading & Writing Assistance Utilizing OCR + DistilBERT for text coherence improvement from a handwritten image

Challenges we ran into

  1. Creating a model for text coherence.
  2. Improving accuracy of the Gen AI models.
  3. Making those models context-aware.
  4. Deploying

Accomplishments that we're proud of

  1. By creating this project, we were able to help the people with dyslexia.
  2. Helping the dyslexic individuals with their day to day activities.
  3. Improving their memory, spellings and creating phonological awareness through gamification techniques.
  4. Creating the perfect UI with dyslexic fonts.

What we learned

  1. Deployment
  2. How to use RAGs for making the LLM models context-aware.

What's next for LexiEaseAI - An AI Powered Dyslexic Support System

  1. Publishing a research paper and file a patent for our project.
  2. Aiming to connect with NGOs.

Built With

Share this project:

Updates