DysgraphAI

A playground using machine learning to detect dysgraphia in handwriting, with an interactive demo to help raise awareness about it.

What is Dysgraphia?

“Dysgraphia is a deficiency in the ability to write, primarily handwriting, but also coherence. Dysgraphia is a transcription disability, meaning that it is a writing disorder associated with impaired handwriting, orthographic coding, and finger sequencing (the movement of muscles required to write). It often overlaps with other learning disabilities such as speech impairment, attention deficit disorder, or developmental coordination disorder.”

^ From Wikipedia

 The Dataset

The dataset for the machine learning model used in the playground was created originally, by digitally drawing about 50 samples of normal handwriting and handwriting with dysgraphia.

 Creating and Training The Model

The model was created and trained using CreateML, using the interactive image classification tool; and the console output for further tuning.

 Credits and Thanks

DSF for the explanation of dysgraphia featured in the playground Sasmito Adibowo for his blog post on how to add .mlmodel files to Swift Playground

Built With

  • animation
  • core-create-ml-core-core-image
  • ml
  • swift
  • uikit
Share this project:

Updates