Inklusive - Dyslexia detection via handwriting analysis

• 

  main page

• 

  quick glance at our process

Inspiration

From the beginning, we knew we wanted to explore and implement machine learning in this project. Initially, we brainstormed many ideas from weight lifting to makeup, however, eventually, we settled on an app that could detect dyslexia through a ML model. Our inspiration stemmed from our combined interest in Biology, Cognitive Science, and education, and how these fields together can be explored through Machine learning. Our constant focus was creating an application or tool to serve the community. We utilized these passions along with our team’s ML skills to create an app that would be able to detect, on a large scale, a common symptom of dyslexia: letter reversals in handwriting. We also found a data set for this project which was another major motivation to pursue this application of ML.

What it does

Discovers early symptoms of dyslexia by identifying whether handwriting is written by a person or not, based on the common symptom of letter reversals, providing them a doctor-free detection. This web app can be used in large classrooms to decrease the time spent by teachers on analyzing handwriting.

Step 1: User uploads an image. Step 2: Keras OCR locates text in the image. Step 3: Text is cropped and isolated. Step 4: Dyslexic vs. non-dyslexic model classifies the text. Step 5: Data for training: procedurally combined letter collections. Step 6: Multiple models for different sizes. Step 7: Closest aspect ratio model is used for classification.

How we built it

In terms of front-end development, we used pure HTML and CSS and for the backend, we used Django and Python for our machine learning model.

Challenges we ran into

Bridging between static and non-static servers due to GitHub not supporting non-static servers and Getting the machine learning algorithm to work alongside our website together

Accomplishments that we're proud of

We are proud of being able to produce this project in a minimal amount of time, collaborate with one another, and also learn skills from each other while creating this project, it’s truly been a learning experience.

What we learned

Collaborating despite our differences in time zones meant that we had to compensate for each other's schedules. We think this helped boost overall communication and teamwork and taught us how to work remotely with others.

What's next for Inklusive - Dyslexia detection via handwriting analysis

The next steps for this app involve continuous improvement through algorithm and model updates to enhance dyslexia detection. Expanded features such as voice recognition and multi-language support will be introduced, while partnerships with educational institutions and dyslexia organizations will broaden its impact. Data privacy measures will be strengthened to ensure user trust and protection of sensitive information. AI-driven interventions and personalized learning paths will offer targeted support for dyslexic learners. User feedback will be collected to address specific needs, and efforts will be made to make the app globally accessible in various regions, languages, and devices.

Built With

• css
• django
• figma
• html
• javascript
• kera-ocr
• python